Alcohol & Alcoholism Vol. 41, No. 2, pp. 181–187, 2006 doi:10.

1093/alcalc/agl001
Advance Access publication 2 February 2006

ALQOL 9 FOR MEASURING QUALITY OF LIFE IN ALCOHOL DEPENDENCE

LAURENT MALET1,5*, PIERRE-MICHEL LLORCA1, BÉRÉNICE BERINGUIER2, PHILIPPE LEHERT3,4 and
BRUNO FALISSARD5
1
Centre Hospitalier Universitaire, Psychiatrie B, rue Montalembert BP 69 and 2Cente Hospitalier Sainte-Marie, 10 avenue Franklin Roosevelt,
F-63003 Clermont-Ferrand Cedex 1, France, 3Department of statistics, Faculty of Economy, FUCAM, MONS, Belgium, 4Faculty of medicine,
University of Melbourne, Melbourne, Australia and 5INSERM U669—Hôpital Cochin, Maison de Solenn, 97 Boulevard du Port-Royal,
F-75679 Paris cedex 14, France

(Received 12 May 2005; first review notified 12 July 2005; in revised form 28 December 2005; accepted 4 January 2006;
advance access publication 2 February 2006)

Abstract — Aims: Quality of life (QoL) is an important factor of outcome tracking and treatment in alcohol misuse. A 9-item QoL
scale, AlQoL 9, obtained from the generic SF 36, is proposed as a measure that characterizes the QoL of alcohol-dependent patients.
Our objective was to study the psychometric properties of this subscale. Methods: AlQoL 9 was evaluated in two study groups of

Downloaded from https://academic.oup.com/alcalc/article/41/2/181/135080 by guest on 10 April 2024

patients with DSM-IV diagnosis of dependence: 104 inpatients, and 114 outpatients. Severity of dependence, alcohol consumption,
psychiatric, and somatic comorbidities were assessed. We studied the global properties of AlQoL 9 and its structure. Results: Cronbach
a-coefficients in both populations indicated good internal consistency (0.71 and 0.85). Test–retest intraclass coefficients for a 2-day
interval in hospital were in the range 0.57–0.78. Principal component analysis found a unidimensional scale. This subscale has proper-
ties that are consistent with the concept of QoL in alcohol dependence, i.e. lowered QoL compared with the general population, influ-
enced by gender, and depression. Conclusions: AlQoL 9 epitomizes QoL in alcohol-dependence. It gives a global measurement with
good psychometric properties. It could be used in clinical practice as a diagnosis and management support instrument and may also
be useful in research for evaluating treatment efficacy.

INTRODUCTION eliminating non-relevant and (or) redundant items (strictly-

Health-related quality of life (QoL) has only been a recognized
medical variable for 20 years. Its utility has been gradually
acknowledged, especially in chronic health disorders. QoL
has to be considered when making decisions about healthcare,
because improvement of QoL is not an automatic result of
improved clinical status.
In the psychometric tradition, the concept of QoL comprises
several dimensions. Three areas seem to be essential: social,
psychological, and physical (Testa and Simonson, 1996). It
is generally accepted that the more dimensions covered in
the questionnaire, the richer the information obtained, giving
more universally applicable results. Multi-dimensional health
status measures give health-related QoL profiles (sets of
standardized scores). Two different types of scales have to
be considered: generic, and disease-specific scales.
Generic, and disease-specific scales
Generic scales can address health-related QoL adequately in
all diseases. They are useful in comparing QoL or effects of
treatment between quite different diseases. Their universality
gives a large amount of data in many pathologies. The Short
Form 36 (SF 36) is one of the most widely used and simplest
generic scales. SF 36 has been validated in a French-language
version, and a reference manual gives values for the general
French population (Perneger et al., 1995).
Disease-specific QoL scales are optimized in term of
sensitivity and specificity for a particular pathology. Each
disease does not alter all QoL dimensions, and items
that are not influenced by the disease dilute the other
items that do change and (or) are altered. Most disease spe-
cific scales have been developed from generic scales by
*Author to whom correspondence should be addressed at: Tel.: +33 4 73 75 21
24; Fax: +33 4 73 75 21 29; E-mail: lmalet@chu-clermontferrand.fr
linked items with identical evolution). Specific QoL scales
do not allow comparison between diseases but give a sharp
and precise measurement for the considered disease as a
complement to the clinical assessment. They are constructed
in such a way as to give a high sensitivity to change.
QoL in alcohol dependence
The French Alcohology Society (SFA) consensus conference
on ‘support to the alcohol-dependent subject following
detoxification’ issued the target of helping patients to recover
a good QoL, and emphasized both the lack of information
material in this field and the absence of a specific measure-
ment tool (Agence Nationale d’Accréditation et d’Evaluation
en Santé, 1999).
QoL is in fact an essential indicator in this multifactorial
pathology for both the diagnostic, and the therapeutic stages.
The DSM-IV definition of dependence does in fact deal with
these aspects to some extent, since five of the nine items
cover the social, familial, and occupational consequences
of alcohol consumption (the other tolerance and craving
items are more closely related to the quantification of
consumption).
Foster et al. (1999) reviewed ‘QoL in alcohol-dependent
subjects’. They noted the ‘paucity of papers’ and listed
24 publications on the topic from 1982 to 1997. A Medline
BIDS database record search from 1998 to 2004 with
the same key words identifies only three additional papers.
QoL of alcohol-dependent subjects is reduced compared
with that of a normative healthy population (Hunt and
McEwen, 1980; Welsh et al., 1993, McKenna et al., 1996),
with differences between gender (Qol poorer for women)
(Foster et al., 2000). As regards SF 36, the role limitation
and psychological functioning scores are lower than those of
physical and functioning dimensions (Daeppen et al., 1998),
and handicap is most important in physical and emotional

181

The Author 2006. Published by Oxford University Press on behalf of the Medical Council on Alcohol. All rights reserved
182
Table 1. Structure of SF 36 and items of AlQoL 9
Dimensions of SF 36 Number of items Range of scale
PF (physical functioning)* 10 1–3
BP (bodily pain)* 2 1–6
GH (general health)* 5 1–5
RP (role physical/limitation)* 4 1–2
MH (mental health)** 5 1–6
RE (role emotional/limitation)** 3 1–2
VT (vitality)** 4 1–6
SF (social functioning)** 2 1–5
Dimensions of SF 36 are scored on a 0–100 scale, with higher scores indicating better health functioning.
*Items of the SF 36 Physical Health Dimension.
**Items of the SF 36 Mental Health Dimension.
role limitations, and social functioning (Morgan et al., 2004).
Psychiatric comorbidities—especially depression—(Beattie
et al., 1993; Daeppen et al., 1998; Driessen et al., 1998), dis-
turbed sleep (Foster et al., 2002), social and other alcohol-
related problems (Patience et al., 1996) are major factors
linked to QoL.
A subscale of SF 36 characterizing alcohol dependence
SF 36 is a generic scale derived from an observational study
(Brazier et al., 1992). Its internal and external validity in
alcohol-dependent populations is established (McKenna et al.,
1996, Patience et al., 1997). It is a self-questionnaire that
yields profiles, and uses Likert-type scales. It comprises eight
dimensions: physical functioning (PH), bodily pain (BP),
mental health (MH), energy/vitality (VT), general health
perception (GH), role limitations due to physical problems
(RP), role limitations due to emotional problems (RE), and
social functioning (SF) (Table 1).
The NEAT study provided QoL data in two groups after
detoxification. QoL was measured with SF 36 at M0, M3,
and M6 (Pelc et al., 2002; Morgan et al., 2004). The objective
was to identify the best subscale of SF 36 specific for alcohol-
dependence defined as the smallest number of items covering
a set of non-redundant dimensions satisfying a reasonable
reliability (Lehert, 2002; Lehert and Poldrougo, 2002). At
M3 and M6, change in QoL was shown to be essentially influ-
enced by cumulative abstinence duration (CAD) (Morgan
et al., 2004). Thus, for reasons of simplicity, sensitivity was
estimated by rank correlation with CAD. The following iterat-
ive procedure was used: (i) at start, select the most sensitive
dimension D1 and identify the kernel K1 of D1, the smallest
number of items such that a > 0.7; (ii) from this first dimen-
sion, select the dimension D2 defined as the most sensitive
and least correlated with any already selected dimension,
and identify its kernel K2; and (iii) iterate while all dimensions
are entered or no remaining dimension is such that its
minimum correlation with an existing dimension is > 0.7. As
the algorithm is based on iterative procedures, results may
vary between samples. Bootstrap was used, first in using the
whole sample, and then in excluding subsets of patients char-
acterized by various socio-cultural or alcoholic characteristics
to assess invariance to unspecific factors. The number of
resulting dimensions was found to be quite stable (index of
stability = 93%) and invariably reported role, mental, and
physical dimensions (but not in that order). However the
choice of items in the kernel was not stable, so as the most
L. MALET et al.
Items of alcohol SF 9
PF4 ‘are you limited in climbing several flights of stairs?’
BP1 ‘how much bodily pain have you had in the last 4 weeks’
GH1 ‘in general would you say your health is excellent/ . . . /poor?’
RP4 ‘were you limited in the kind of work or other activities?’
MH1 ‘have you been a very nervous person in the last 4 weeks?’
MH4 ‘Have you felt downhearted or low in the last 4 weeks?’
RE2 ‘have you accomplished less than you would like in your work/activities?’
VT3 ‘did you feel worn out in the last 4 weeks?’
SF2 ‘have problems interfered with your normal social activities?’
stable results appeared to be at least six, and at the most
Downloaded from https://academic.oup.com/alcalc/article/41/2/181/135080 by guest on 10 April 2024
nine, this last option was retained, which consists of the items
shown in Table 1 and which we call AlQoL9.
Objectives
We considered AlQoL 9 as a scale characteristic of alcohol
dependence-related QoL and set out to validate its French ver-
sion, using the parent French validated SF 36 (Leplège et al.,
2001). Our study had two objectives: (i) structural analysis
of AlQoL 9 (to assess number of dimensions), and (ii) to
assess sensitivity of AlQoL 9, focusing on whether it was
able to highlight all known data on the QoL of dependent
patients. We studied separately two alcohol-dependent patient
populations (inpatients and outpatients) starting treatment to
test the reproducibility of the results.
MATERIALS AND METHODS
Study populations
Subjects were patients aged between 18 and 65, seeking treat-
ment, who met the DSM IV criteria for alcohol dependence,
and all subjects gave their informed consent. General exclu-
sion criteria were patients for whom alcohol dependence
was not the main diagnosis on axis I of DSM IV (mental
retardation, schizophrenia, or other psychotic disorder, bipolar
mood disorder), any other addiction (except tobacco), and
severe personality disorders (in particular psychopathic and
borderline patients). Subjects with anxiety or depression (as
a secondary diagnosis) were not, however, excluded.
Inpatients were recruited in specialized addiction treatment
wards (the University Hospital and the Sainte-Marie Hospital
in Clermont-Ferrand). They were included on admission
from March to September 2002. The diagnosis was made by
clinicians specialized in alcohol-related pathologies and who
were familiar with DSM IV.
Outpatients were included from March to December 2002
by 30 general practitioners. These GPs were chosen at random
from the official public list of physicians in Central France.
They attended training sessions for the mini international
neuropsychiatric interview (MINI) to diagnose DSM-IV
dependence and anxiety and mood disorders.
Data and measurements
The questionnaires including AlQoL 9 were filled out by
the patients on the day they were admitted to hospital or at
 ALQOL 9 FOR MEASURING QUALITY OF LIFE
consultation. The QoL questionnaire was filled out a
second time (as a re-test) by inpatients 48–72 h after their
hospitalization (the QoL questionnaire looks at the previous
month and does not identify any differences at a 2- to 3-day
interval).
Alcohol-dependence is a complex multifactorial pathology,
and patients are heterogeneous. There is no real consensus on
definition of subgroups or on how to describe the patients. In
addition, many factors are able to influence QoL. Therefore,
we collected numerous descriptive data in various fields we
considered relevant: sociodemography, alcohology, psychi-
atry, and somatics.
 The usual sociodemographic data collected were gender,
age, marital status, living arrangements, and employment
status
 Prior alcohol history. Two datasets were collected: duration
of alcohol misuse (<2, 3–5, 6–10, and >10 years) and earlier
detoxification (none, 1–5, >5 years).
 Severity of alcohol-dependence

Psychiatric comorbidities
 Somatic comorbidities
We did not use the SADQ (severity of alcohol-dependence
questionnaire) or ADS (alcohol-dependence scale) severity
scales since they have not been validated in French.
DSM-IV diagnoses alcohol-dependence as positive scores
for at least three of the nine diagnostic criteria for the last
12 months. Items 1–4 identify physical dependence. Severity
of dependence was based on summation of DSM IV criteria
and simplified by division into three categories: moderate,
mild, and severe for 3–4, 5–6, and 7–9 positively-scored
criteria, respectively (Woody et al., 1993).
The severity of the alcohol-dependence was also evaluated
by clinicians using clinical global impression (CGI) which
scores from 0 (not ill) to 7 (extremely severe).
Alcohol consumption can be used as an index of severity.
Alcohol consumption was measured by the first three items
of the AUDIT questionnaire indicating number of drinks
per day of consumption, number of days of consumption
per week, and number of 5-plus drinks on occasion per month
(Gmel et al., 2001). AUDIT consumption scores go from 0
(abstinence) to 15 (more than seven drinks per day, each day).
Psychiatric comorbidities. Anxiety and mood co-
morbidities were accurately evaluated with the hospital
anxiety and depression (HAD) scale in inpatients, and MINI
interview for outpatients. HAD gave anxiety and depression
scores. The common threshold usually retained to assert the
presence of each disorder is eight (Herrmann, 1997). MINI
is a standardized diagnostic tool consistent with DSM criteria
(Sheehan et al., 1998). It was used here to diagnose major
depressive episodes and anxiety disorders.
Somatic comorbidities. Alcohol-related somatic comorbid-
ities were identified by the clinicians. They are grouped
into six general categories: hepato–pancreatic (e.g. hepatitis,
steatosis, or cirrhosis,), gastric (hemorrhagic gastritis, and
the like), cardiac (hypertension, cardiomyopathy, and the
like), neurological (polyneuritis, Korsakov’s syndrome, and
the like), oncological, or trauma-related. We developed a
measure of physical comorbidities based on a simple
count. Our index of somatic comorbidities scores from 0
183
(no comorbidity) to 6 (at least one comorbidity in each
category).
Analysis
We separately conducted two identical analyses in each of the
two populations to demonstrate any reproducibility (or to help
explain any characteristics-related differences between the
populations).
 Both populations were described and compared using all the
data collected.
 General properties of AlQoL 9:
- Distribution of responses per item was studied to assess
acceptability (missing items) and upper and lower limit
effects of the scale.
Downloaded from https://academic.oup.com/alcalc/article/41/2/181/135080 by guest on 10 April 2024
- Assessment of internal consistency using Cronbach’s a-
coefficient for global scale. Significance of internal con-
sistency were interpreted according to Cicchetti
(Cicchetti, 1994): <0.70 was classified as unacceptable,
between 0.70 and 0.79 as fair, between 0.80 and 0.89 as
good, and >0.90 as excellent.
- Test–retest reliability of the AlQoL 9 scale was evaluated
with the intraclass correlation coefficients (ICC) for inpa-
tients. Qualitative interpretations of the ICC were based
on the recommended ranges: ICC <0.40 was poor, 0.40–
0.59 was fair, 0.60–0.74 was good, and 0.75–1 was excel-
lent (Cicchetti, 1994).
 Structural analysis consisted in the following.
- Screeplot to characterize the number of dimensions con-
stituting the scale and exploratory principal dimension
analysis (PCA) to represent inter-item correlations graph-
ically.
- Internal consistencies for each dimension possibly identi-
fied, and Pearson’s correlation coefficients between items
and dimensions. Pearson’s coefficients were interpreted
according to Cohen’s definitions of the size effect (Cohen,
1992): small when 0.10 < r < 0.30, medium when 0.30
< r < 0.50, and large when 0.50 > r (Cohen, 1992).
 Informative properties (sensitivity and specificity) of
AlQoL 9 consisted in evaluating the ability of the scale to
find known QoL data in alcohol-dependent patients.
Scores for each item were compared with data from the
French general population and were compared between
gender. A multivariate linear regression model was
built using sociodemographic data (gender, age, living
arrangement, and employment), severity of dependence
according to DSM IV, alcohol consumption, psychiatric,
and somatic comorbidities, earlier detoxification and age
of alcohol misuse as adjustment variables. All collected
data were considered as potential predictors of QoL. Each
variable was considered as binary, nominal, or ordered as
appropriate
Data analysis was carried out using SAS software. The
statistical tests used are stated in the text. All statistical tests
used a two-sided risk a of 5%. The graphic results (screeplot
and PCA) were obtained using the R statistical package
using the ‘psy’ library (www.r-project.org).
184 L. MALET et al.
Table 2. Clinical data General properties of AlQoL 9
Inpatients (%) Outpatients (%) The validity of the AlQoL 9 instrument was excellent. There
(N = 104) (N = 114) Statistics (P) was only one missing data point (a non-response to item PF4
Physical dependence 84 83 0.57 in outpatients). There was no upper limit effect for the global
Severity of dependence score in either of the two populations. There was a negligible
Moderate 47 41 lower limit effect in the outpatients, which was calculated as
(3 or 4 items) (%) 1.8% (0% of inpatients).
Mild (5 or 6 items) 36 39.5
Severe (7, 8 or 9 items) 18 19.5 The internal consistency of the AlQoL 9 was high, with a
Cronbach’s a-coefficient of 0.85 in outpatients and 0.71 in
Age of alcohol misuse 0.74
<2 years 8 5
inpatients.
3–5 years 14 15 Test–retest intraclass coefficients (ICC) for a 2-day interval
6–10 years 17 20 in hospital were in the range 0.57–0.78 for the items: 0.78 for
>10 years 61 60 GH1, 0.73 for SF2, 0.72 for MH4, 0.67 for PF4, 0.64 for MH1,
Earlier weaning 0.001 0.64 for VT3, 0.60 for BP1, 0.57 for RP4, and 0.57 for RE2.
None 29 54 ICC was 0.81 for the global AlQoL 9 score.

Downloaded from https://academic.oup.com/alcalc/article/41/2/181/135080 by guest on 10 April 2024

1–5 years 60 42
>5 years 11 4
Alcohol consumption 14.5 (3.8) 13.4 (4.5) 0.004 Structural analysis
[mean (SD)] A screeplot with simulations (available on request) showed
Clinical global impression <0.0001 a strong first dimension of AlQoL 9 in both populations, and
Not ill 0 5 suggested a weak second dimension in the inpatients.
Very mild or mild 4 29 PCA showed fairly similar diagrams (Fig. 1). The mental
Moderate 37 15
Marked or severe 57 51
and physical health items were quite well distinguished and
Extremely severe 2 0 relatively well aggregated. Item RE2 came out as the less
‘stable’. The vitality item (VT3) was closest to the global score.
Anxiety (HAD) 76 — 0.11
Anxiety (MINI) — 84 These graphical results may lead to group three physical health
Depression (HAD) 44 — 0.08 items (GH1, BP1, and PF4) in a physical health dimension
Depression (MINI) — 55 (PHD) and the other six items (five mental health items and the
Somatic comorbidities 0.55 (0.46) 0.53 (0.60) 0.18 RP4 role physical/limitation) in a mental health dimension
[mean (SD)]
(MHD).
MINI is the mini international neuropsychiatric interview used to detect Inter-dimension correlations between PHD and MHD
anxiety or mood disorders in outpatients. HAD is the hospital anxiety and were strong in outpatients (r = 0.61) and medium in inpatients
depression scale used for inpatients. Alcohol consumption is based on (r = 0.31). The correlations between each item and its dimension
AUDIT questions and scored from 0 (abstinence) to 15 (>7 drinks per were in the range 0.43–0.88. The correlations between each item
day, each day). Statistics are P-values of Mann–Whitney tests for compar-
ing means and exact Fisher tests for comparing percentages. and the other dimensions were in the range 0.07–0.57 (Table 3).
For both dimensions, Cronbach’s a-coefficients ranged
from 0.46 to 0.82.
Structural analysis is not indicative of a true 2D scale struc-
ture, and thus argues in favour of using a simple global score
RESULTS
for AlQoL 9, theoretically ranging from 9 (lowest QoL) to
41 (optimum QoL). The mean score obtained over the SF
Population descriptions
36 reference sample in the general population is 32.
The two populations were of similar sizes—104 inpatients
and 114 outpatients—and were sociodemographically com-
parable. The sex ratio (male:female) was 4:1 (82 and 83%, Informative properties
respectively). Mean ages were 44 (9.6) and 48 (9.1) years, Based on the SF 36 reference data for the French general
respectively. In both populations most of the patients lived population, the QoL of alcohol-dependent patients was found
alone (58 and 56%), and half of them were unemployed (44 to be significantly poorer than that of the general population
and 53%). (P < 0.0001 by Hotelling’s T2, which allows all the items to
Clinical characteristics were more heterogeneous between be compared simultaneously). Two physical health items
samples (Table 2). Slightly >40% presented moderate depend- (PF4 and BP1) were closest to the general population. Three
ence, and >80% presented a physical dependence according mental health items (MH1, MH4, and RE2) were found to
to DSM-IV criteria. Nearly 90% of the inpatients (only 57% have deteriorated most (Fig. 2).
of outpatients) were considered moderately or markedly ill In addition, QoL was more severely impaired in the women
according to CGI. 40% of the outpatients were considered than in the men, with respective global scores of 19.8 (±5.0)
normal, borderline, or mildly ill compared with only 4% of vs 24.3 (±3.6) in outpatients (Mann–Whitney P = 0.012) and
the inpatients. Almost three-quarters of inpatients had an 21.8 (±3.6) vs 25.7 (±5.1) in inpatients (Mann–Whitney
experience of prior detoxification (46% of outpatients) and P = 0.002).
their alcohol consumption was significantly greater. There The regression model found that gender and depression
were no between-population differences in psychiatric had a significant impact on QoL score in both inpatient and
(depression or anxiety) or somatic comorbidities. outpatient populations. Other significant covariates were
 ALQOL 9 FOR MEASURING QUALITY OF LIFE 185

x = F1 : 33% var x = F1 : 50% var

y = F2 : 16% var y = F2 : 11% var

mh1
mh4 mh1
sf2 re2
mh4
sf2
rp4
QoL
rp4 QoL
vt3
vt3

bp1 re2 gh1

pf4

gh1 bp1

Downloaded from https://academic.oup.com/alcalc/article/41/2/181/135080 by guest on 10 April 2024

pf4

Fig. 1. PCA according to the first two factors (F1, and F2). These two factors represent 49% of variance in inpatients (left) and 61% in outpatients (right). The two
figures are relatively similar. Items GH1, BP1, and PF4 seem to correspond to a PHD, the other items to a MHD. QoL indicates the global score of AlQoL 9.

Table 3. Hypothesis of a MHD and a PHD consolidating three and six people are unemployed, and only 30% live alone (Chaleix,
items, respectively 2001; Aerts and Bigot, 2002).
Correlations with PHD Correlations with MHD Inpatients seemed to be more severely affected. Very few
had CGI scores of very mild or mild (4 vs 29% for outpa-
Items Inpatient Outpatient Inpatient Outpatient tients). Their alcohol consumption was higher, and they had
PHD? PF4 0.65 0.76 0.16 0.49 more often experienced earlier detoxifications. There were
BP1 0.86 0.88 0.31 0.48 no significant differences between the two population groups
GH1 0.59 0.80 0.20 0.54* in terms of anxious (75% of cases), depressive (50% of
MHD? RP4 0.22 0.50* 0.57 0.65 patients), or somatic comorbidities. It would appear that to
MH1 0.07 0.40 0.69 0.78
MH4 0.23 0.49 0.81 0.88 some extent, patients generally enter (over the long-term?)
VT3 0.35 0.57* 0.68 0.79 into a defined hospital or outpatient healthcare system.
RE2 0.31 0.39 0.43* 0.62 Patients with a history of hospitalizations have probably
SF2 0.27 0.46 0.71 0.76 been through more detoxifications.
MHD 0.31 0.61* * *
The general properties of AlQoL 9 seem to be satisfactory.
*Indicates correlations in opposition of this 2D hypothesis: too week correla- Acceptability was excellent. Internal consistency was ‘fair’ for
tion between one item and its too weak component, or too strong correlation inpatients (0.71) and ‘good’ for outpatients (0.85).
between one item and the other component. Test–retest reliability was ‘excellent’ for global score. It
was ‘excellent’ for the general health item (GH1), ‘good’ for
alcohol consumption and anxiety for inpatients, and somatic six items, and ‘fair’ for the two role limitation items (RP4
comorbidities for outpatients (Table 4). and RE2). These results only indicate good test–retest
reliability, which was only carried out in one of the two
populations at a time when detoxification difficulty is
DISCUSSION highest, (i.e. 48–72 h after detoxification was started). In
fact, 12% of patients presenting with major signs of with-
This study was constructed in two independent patient drawal were deemed unfit to fill out the questionnaire a second
populations enrolled in relatively small numbers (104 and time.
114 patients). The separate per-population analyses (and Structural analysis in each of the two populations globally
reproducibility of results) enhance its validity. favours a one-dimensional scale. PCA showed a consolida-
A sex ratio of 4–5:1 is normal in alcohology studies with tion of items derived from the MHD and the PHD of SF 36,
clinical samples (Reynaud and Parquet, 1999). The two respectively. An aggregation of these items into two dimen-
populations were similar. Mean ages of 44 and 48 years sions would give internal consistency for each dimension
are consistent with the time taken to establish alcohol from 0.82 (good) to 0.46 (unacceptable). Correlations between
dependence (60% of patients had alcohol misuse since the items derived from one dimension and the items of the
over 10 years). The high proportion of patients living alone other dimension were very often ‘large’ (>0.50), whereas
(56–58%) and unemployed (close to 50%) reflects the social they should be ‘small’ for separate dimensions.
and familial causes or consequences of alcohol dependence. Despite this reductive unidimensional approach, AlQoL 9
In the general French adult population (>18 years), 9% of possesses excellent informative qualities. The scale showed
186 L. MALET et al.

100%
93%
90%
84%
81% 80% 79% 78%
77%
80% 74% 74% 73% 74% 76% 74%
72%
70% 68%
66% 67% 66%

60%

40%

20%

0%

9
1

3
4

P4

P1

E2

oL
PF

VT

SF
H

H
R

R
M

M
G

Downloaded from https://academic.oup.com/alcalc/article/41/2/181/135080 by guest on 10 April 2024

lQ
A
in-patients out-patients

Fig. 2. QoL scores for each item and global scale in the two populations compared with the French general population. 100% indicates the general population
reference level.

Table 4. Covariates of a multiple regression model to explain the global comorbidities has more ravaging effects in outpatients, who
AlQoL 9 score (socio-demographic and clinical variables)
are managed less intensively. The index of somatic complica-
Inpatients (P-value) Outpatients (P-value) tions used as an indicator in this study remains limited.
Gender 0.0462* 0.0099*
Age 0.8439 0.7390
Living arrangements 0.3176 0.5584 CONCLUSION
Employment 0.0653 0.3371
DSM-IV severity 0.2586 0.2842
Age of alcohol misuse 0.5589 0.1246 AlQoL 9 is a health-related QoL scale which epitomizes QoL in
Earlier weaning 0.2906 0.6323 alcohol-dependence. The nine items that compose it (derived
Alcohol consumption 0.1889 0.0456* from the SF 36 generic scale) are highly characteristic of
Anxiety 0.2194 0.0002*
Depression 0.0264* 0.0081*
those aspects of QoL most affected by alcohol dependence. It
Somatic comorbidities 0.0048* 0.0987 demonstrates excellent informative qualities, and is sensitive
to most of the factors known to be involved in the QoL of
*Indicates significant (P < 0.05) variables. This model identified two known alcohol-dependent persons.
variables: gender, and depression. This very specific instrument could probably find appli-
cations in medium or long-term treatment evaluation. The
good specificity for alcohol dependence, and in both popula- NEAT princeps study showed a very good sensitivity to
tions it was able to identify the following: change at 3 and 6 months, but further investigations will be
needed.
- a poorer QoL than in the general population; AlQoL 9 also has an immediate clinical utility, as it is
- more important deficit in mental health than in the physical easy use and high specificity should help patients in denial
dimension; to become more aware of their condition by underlining—
- a poorer QoL in women than in men (considering global and quantifying—the impact of their alcohol consumption on
score or each item); and their daily lives. This instrument could provide a complement
- depression as a major factor of QoL. to motivational counselling and facilitate the change process.
It may thus help practitioners both to broaden their appraisals,
QoL in inpatients appears not only to be related to depres-
which are currently based solely on alcohol consumption, and
sion but also to anxiety and alcohol consumption. Although
improve follow-up, as QoL may be a predictor of relapse.
depression and anxiety disorders were a little less frequent
in the inpatient population than the outpatients, they were Acknowledgements — This study, which was financed by a grant from the
probably more severe, thus causing deterioration of QoL that Programme Hospitalier de Recherche Clinique (PHRC) National 2001, was
initially led to the hospitalization. It is difficult to explain the promoted by the Clermont–Ferrand University Hospital Center.
relationship between QoL and somatic comorbidity that was
identified in the outpatients. One explanation would be that
the outpatients did not experience more somatic problems REFERENCES
but more serious somatic problems evolved to a stage
affecting QoL. Indeed, over half of these patients had never Aerts, A.-T. and Bigot, J.-F. (2002) [National survey about employ-
detoxified and presented long-standing, uninterrupted alcohol ment in 2002], National Institute of Statistics and Ecnomical
consumption. It may also be that the subjective experience of Studies, INSEE Première, No. 857.
 ALQOL 9 FOR MEASURING QUALITY OF LIFE
Agence Nationale d’Accréditation et d’Evaluation en Santé (1999)
[Consensus conference in alcohology: objectives, indications
and modalities for detoxification the alcohol dependent patient].
Alcoologie addictologie tome 23, 109–388.
Beattie, M. C., Longabaugh, R., Elliott, G. et al. (1993) Effect of the
social environment on alcohol involvement and subjective
well-being prior to alcoholism treatment. Journal of Studies on
Alcohol 54, 283–296.
Brazier, J. E., Harper, R., Jones, N. M. et al. (1992) Validating the
SF-36 health survey questionnaire: new outcome measure for
primary care. British Medical Journal 305, 160–164.
Cicchetti, D. V. (1994) Guidelines, criteria, and rules of thumb for
evaluating normed and standardized assessment instrumentsin
psychology. Psychological Assessment 6, 284–290.
Chaleix, M. (2001) [7,4 millions adults live alone in 1999]. National
Institute of Statistics ans Economical Studies, INSEE Première,
N 788.
Cohen, J. (1992) A power primer. Psychological Bulletin 112,
155–159.
Daeppen, J. B., Krieg, M. A., Burnand, B. et al. (1998) MOS-SF-36
in evaluating health-related quality of life in alcohol-dependent
patients. American Journal of Drug and Alcohol Abuse 24,
685–694.
Driessen, M., Veltrup, C., Weber, J. et al. (1998) Psychiatric
co-morbidity, suicidal behaviour and suicidal ideation in alcohol-
ics seeking treatment. Addiction 93, 889–894.
Foster, J. H., Powell, J. E., Marshall, E. J. et al. (1999) Quality
of life in alcohol-dependent subjects—a review. Quality of Life
Research 8, 255–261.
Foster, J. H., Peters, T. J. and Marshall, E. J. (2000) Quality of life
measures and outcome in alcohol-dependent men and women.
Alcohol 22, 45–52.
Foster, J. H., Peters, T. J. and Kind, P. (2002) Quality of life, sleep,
mood and alcohol consumption: a complex interaction. Addiction
Biology 7, 55–65.
Gmel, G., Heeb, J. L. and Rehm, J. (2001) Is frequency of drinking an
indicator of problem drinking? A psychometric analysis of a
modified version of the alcohol use disorders identification test
in Switzerland. Drug and Alcohol Dependence 64, 151–163.
Herrmann, C. (1997) International experiences with the hospital
anxiety and depression scale—a review of validation data and
clinical results. Journal of Psychosomatic Research 42, 17–41.
Hunt, S. M. and Mcewen, J. (1980) The development of a subjective
health indicator. Sociology of Health and Illness 2, 231–246.
Lehert, P. (2002) The measurement of quality of life in alcoholism
and assessment of the efficacy of Campral. In Proceedings of
187
the Fourth European Symposium on Acomprosate in Seville,
March 2001, Seville, Vol. 2, N 1.
Lehert, P. and Poldrougo, F. (2002) Acomprosate and quality of life.
In Proceedings of the Symposium of International Society on
Addiction Management ISAM, April 2002, Trieste, Vol. 1, N 1.
Leplège, A., Ecosse, E., Pouchot, J., Coste, J. and Pernegger, T.
(2001) [The MOS SF 36 questionnaire: French Users Manual].
Estem eds, Paris.
McKenna, M., Chick, J., Buxton, M. et al. (1996) The SECCAT
survey:I. The cost and consequences of alcoholism. Alcohol and
Alcoholism 31, 565–576.
Morgan, M. Y., Landron, F. and Lehert, P. (2004) Improvement
in quality of life after treatment for alcohol dependence with
acamprosate and psychosocial support. Alcoholism: Clinical and
Experimental Research 28, 64–77.
Patience, D., Buxton, M., Chick, J. et al. (1997) The SECCAT
survey: II. The alcohol related problems questionnaire as a proxy
for resources cost and quality of life in alcoholism treatment.
Alcohol and Alcoholism 32, 79–84.
Downloaded from https://academic.oup.com/alcalc/article/41/2/181/135080 by guest on 10 April 2024
Pelc, I., Ansoms, C., Lehert, P. et al. (2002) The European NEAT
program: an integrated approach using acamprosate and psycho-
social support for the prevention of relapse in alcohol-dependent
patients with a statistical modeling of therapy success
prediction. Alcoholism: Clinical and Experimental Research 26,
29–38.
Perneger, T. V., Leplege, A., Etter, J. F. et al. (1995) Validation of
a French-language version of the MOS 36-Item Short Form
Health Survey (SF-36) in young healthy adults. Journal Clinical
Epidemiology 48, 1051–60.
Reynaud, M. and Parquet, J. P. (1999) [Persons in difficulty with
alcohol: use, harmful use and dependance]: proposals. CFES,
eds, Paris.
Sheehan, D. V., Lecrubier, Y., Sheehan, K. H. et al. (1998) The
Mini-International Neuropsychiatric Interview (MINI): the
development and validation of a structured diagnostic psychiatric
interview for DSM-IV and ICD-10. Journal of Clinical Psychiatry
59, 22–33; quiz 34–57.
Testa, M. A. and Simonson, D. C. (1996) Assesment of quality-of-life
outcomes. New England Journal of Medecine 334, 835–840.
Welsh, J. A., Buchsbaum, D. G. and Kaplan, C. B. (1993) Quality of
life of alcoholics and non-alcoholics: does excessive drinking
make a difference in the urban setting? Quality of Life Research
2, 335–340.
Woody, G. E., Cottler, L. B. and Cacciola, J. (1993) Severity of
dependence: data from the DSM-IV field trials. Addiction 88,
1573–1579.