Professional Documents
Culture Documents
Quality of Life Instruments Used in Problem Gambling
Quality of Life Instruments Used in Problem Gambling
A R T I C LE I N FO A B S T R A C T
Keywords: The purpose of this systematic review was to identify the instruments used in original articles to measure quality
Review of life (QOL) or health-related QOL (HRQOL) in gambling-disorder patients and to assess their suitability.
Meta-analysis The systematic literature search to identify QOL/HRQOL instruments used among gambling-disorder patients
Gambling was performed in PubMed, Embase and PsycINFO databases up to November 2018. A meta-analysis was per-
Quality of life
formed to study the effect size of the QOL/HRQOL instruments and gambling outcomes after an intervention.
Harm
Thirty-five studies were included. Seven types of instruments aiming at measuring QOL/HRQOL were
identified. These instruments explored twenty-six domains. The instruments used were not properly validated in
the studies. Most of the clinical trials reported a significant difference in QOL/HRQOL between pre- and post-
intervention. These results were concordant with gambling outcomes but had a smaller effect size than gambling
outcomes.
The currently used general instruments are efficient to measure a significant change after an intervention but
might not evaluate specific areas of health related QOL impacted by gambling disorders
⁎
Corresponding author.
E-mail address: nicolas.bonfils@aphp.fr (N.A. Bonfils).
https://doi.org/10.1016/j.neubiorev.2019.06.040
Received 6 March 2019; Received in revised form 26 June 2019; Accepted 29 June 2019
Available online 02 July 2019
0149-7634/ © 2019 Elsevier Ltd. All rights reserved.
N.A. Bonfils, et al. Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
involving four areas: physical, physical well-being, psychological state “gambling disorder” and “problem gambling”. The search was con-
and social relations (Leplège, 1999). On one hand, “quality of life" ducted to identify original studies that reported the use of QOL/HRQOL
describes a subjective feeling of satisfaction with life in domains of instruments in gambling in any country. Electronic database searches
importance to the subjects. On the other hand, “health-related quality were limited to English-language publications.
of life” reports a subjective perception of the impact of the disease and
its treatment(s) on daily life, physical, psychological and social func- 2.4. Study selection
tioning and well-being (Leidy et al., 1999).
QOL has been discussed in the medical literature since the 1960s The study selection process was comprised of the following two
(Elkinton, 1966). It became increasingly important in heath care to phases. First, titles and abstracts of all potentially relevant publications
measure outcomes beyond morbidity and biological functioning were carefully screened and reviewed for eligibility according to in-
(Karimi and Brazier, 2016). The practice of medicine is often based on clusion and exclusion criteria by two authors (AL and NB). Secondly,
the identification and management of symptoms, while patients' ex- the full texts of studies selected at the first step were obtained and re-
pectations go beyond management of symptoms: they are looking for viewed for eligibility, using the same inclusion and exclusion criteria as
optimal well-being (Luquiens and Aubin, 2014). In addiction, QOL as- in Level 1, by the same researchers. When necessary, we contacted
sessment matches with the treatment goal of enhanced client func- authors of relevant studies to obtain additional information, article
tioning and predict treatment adherence (Laudet, 2011). Moreover, texts, and resolve questions about eligibility. If there were disagree-
participants with addiction at all stages of recovery expressed concerns ments between the two authors in terms of selection of articles at the
about multiple areas of functioning (Laudet et al., 2009). QOL is often abstract level/ article level, discrepancies were resolved by consensus.
included as a secondary endpoint in clinical trials, reflecting patients’ Finally, with the same method, we systematically conducted any search
feelings and functioning and the impact of their health condition be- of the reference list of the included studies.
yond simple symptom assessment (Carr et al., 2001). In 2006, Walker
et al. stressed that there was a lack of evaluated or validated psycho- 2.5. Data extraction
metric instruments assessing QOL in gambling disorder (Walker et al.,
2006) whereas evaluated or validated psychometric instruments as- Data were extracted from the full text of articles by both the two
sessing QOL do exist in substance use disorders (Luquiens et al., 2016; investigators working independently. Data extracted included the fol-
Neale et al., 2016). lowing items: aim, research question, design and setting, sample char-
The purpose of this review was (i) to identify the instruments used acteristics, intervention, instrument used to assess QOL/HRQOL and
in original articles to measure QOL/HRQOL in subjects with a gambling result(s) on QOL/HRQOL. For epidemiological studies, the main out-
disorder in all original articles and then (ii) to assess the suitability of come measure nature and results were extracted. For trials, we ex-
these instruments (reliability, content validity, effect size) in the gam- tracted whether QOL/HRQOL was a primary or secondary endpoint and
bling disorder field. the other endpoint nature and results. The data were summarized in
Tables 2a, 2b, 2c (see in results section).
2. Material and methods Psychometric properties of the instrument (reliability, content va-
lidity, effect size) in gambling population were also extracted. The
The systematic review was conducted independently using the QOL/HRQOL instruments are generally tested for two types of relia-
Preferred Reporting Items for Systematic Reviews and Meta-Analyses bility: (i) the internal consistency measured by Cronbach’s alpha and
(PRISMA) statement guidelines for systematic reviews (Liberati et al., (ii) the test-retest reliability estimated by Pearson correlation coeffi-
2009). Our review project was registered in PROSPERO database. Our cient. The content validity refers to the degree to which an instrument
project began in 2016. Since then, we have updated our review and measures the concept it has to measure. We also studied the effect size
implemented it with a meta-analysis. So we have updated our record in of these instruments compared with the effect size of gambling beha-
PROSPERO database (https://www.crd.york.ac.uk/PROSPERO/# vioral measures (See below: Meta-analysis).
recordDetails). Finally, we analyzed currently used QOL/HRQOL concepts in
gambling-disorder subjects. For this purpose, we identified the life
2.1. Selection criteria domains of these instruments by extracting and examining item con-
cepts, after reviewing all relevant instruments. We have included as
Articles were included if they were (i) original articles aiming at many domains as possible but some may overlap .
assessing QOL/HRQOL in problem or pathological gambling/gambling
disorder subjects and using an instrument designated to measure QOL/ 2.6. Meta-analysis
HRQOL by the authors as an outcome and (ii) if subjects were over 18
years old. We excluded (i) articles with subjects that did not fulfill To assess the ability to detect the change of QOL/HRQOL instru-
problem or pathological gambling/gambling disorder criteria and (ii) ments, we also performed a meta-analysis from selected trials. Three
articles with Parkinson disease subjects. trials were excluded due to a lack of data with which to perform the
meta-analysis (no sample size or lack of documentation on outcome).
2.2. Data sources Each of the gambling and QOL/HRQOL pre/post intervention outcomes
were analyzed by calculating the Cohen’s d effect-size for each trial
The systematic literature searched to identify original studies using with the uncertainty of each result being expressed by their 95% con-
QOL instruments in problem gambling was performed in the following fidence intervals (CI). An effect-size of 0.2 to 0.3 is considered to be a
electronic databases: PsycInfo (Ovid), MEDLINE (using PubMed plat- “small” effect, around 0.5 a “medium” effect, and above 0.8, a “large”
form), and Embase (Ovid). effect (Cohen, 1988). Consistent with convention, a conservative esti-
mate of r = 0.70 was imputed as the pre-post correlation in the cal-
2.3. Literature search strategy culation if it was not reported (Rosenthal, 1991; Wu et al., 2015).
Then we calculated the summarized effect-size of gambling out-
A systematic literature search was conducted among peer-reviewed comes and QOL/HRQOL ones using a random-effects model.
journals from 1950 to November 2018 in these three electronic data- Heterogeneity was assessed by calculating the I2 value. The random-
bases. The search strategy included a combination of the following effects model was chosen because of the between trial heterogeneity.
relevant keywords: “quality of life”, “pathological gambling”, The underlying assumption of a random-effects model is that the true
59
N.A. Bonfils, et al. Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
Table 1
Instruments designated as measuring quality of life in included studies among gambling-disorder patients.
SF-36/SF-12 QOLI WHO-QOL-BREF, EUROHIS-QOL 8-item index Q-LES-Q EQ-5D PWI
SF6 –D/SF-8 EQ-VAS
effect could vary between studies based on characteristics of the study • The Medical Outcomes Study Item Short-Form Health Survey (SF)
population or intervention. Indeed, a wide range of gambling and QOL/ with some different length options for the SF: SF-36 (McHorney
HRQOL outcomes were used throughout the different included studies. et al., 1993; Ware, 2000), SF-12 (Ware et al., 1996), SF-8 (Ware and
A random-effect model is more conservative because it produces wider GlaxoSmithKline, 2001) and SF-6D (Brazier et al., 2002, 1998),
confidence intervals for the effect estimates than a fixed-effects model. • The European Quality of Life Questionnaire (EQ-5D) and the EQ-
Publication bias was assessed using a funnel plot. In case of asym- Visual analogue scale (EQ-VAS) (EuroQol Group, 1990; Scott and
metry, Duval and Tweedie's Trim and Fill method was used to impute Huskisson, 1976),
missing studies. A significance level of p < . 05 (two-tailed) was used • The World Health Organization Quality of Life Assessment
for all analyses (Duval and Tweedie, 2000). Instrument (Szabo and WHOQOL Group, 1996) with some length
Finally, we performed a quality analysis of the included randomized options for the WHO-QOL-BREF and the EUROHIS-QOL 8 item
controlled trials following the “Cochrane Collaboration’s tool for as- index (Schmidt et al., 2006),
sessing risk of bias” (Higgins et al., 2011), which includes five risks of • The Quality of Life Enjoyment and Satisfaction Questionnaire (Q-
bias: 1- selection bias, assessing random sequence generation and al- LES-Q) (Endicott et al., 1993),
location concealment; 2- performance bias, assessing blinding of par- • The Quality of Life Inventory (QOLI) (Frisch et al., 1992),
ticipants and personnel; 3- detection bias, assessing blinding of outcome • The Personal Well Being Index – Adult (PWI-A) (Cummins et al.,
assessment; 4- attrition bias, exploring for incomplete outcome data; 2003).
and 5- reporting bias, searching for selective reporting.
All analyses were conducted using the Comprehensive Meta- The instruments used most frequently were the SF (n = 11), fol-
Analysis statistical program, version 3.0 (Englewood, NJ). HJA and NB lowed by the QOLI (n = 10) and WHOQOL (n = 6) (Table 1). In the last
extracted the data and performed this meta-analysis. three years of the search (i.e. 2016, 2017, 2018), the four following
instruments were used: the QOLI (n = 4), the SF (n = 3), the WHOQOL-
BREF (n = 1) and the Q-LES-Q (n = 1) (Table 1). These instruments
3. Results were initially designed for various purposes and referred to different
concepts. The SF instruments were constructed to measure HRQOL. The
The search yielded 423 citations, published between 1997 and EQ-5D is a standardized measure of health status, commonly used as a
2018. Of these, 221 were ineligible after review of the title and abstract. HRQOL instrument. The other instruments were developed to measure
Thirty-five studies were included in this review. The PRISMA flow QOL. None of these instruments were specifically designed to gambling
diagram of the study selection process (Fig. 1) shows the reasons for disorder.
excluding articles. The thirty-five included articles were published be- The methods and the results of the thirty-five articles are presented
tween 2002 and 2017. We identified six types of instruments aiming at in Tables 2a, 2b, 2c.
measure QOL or HRQOL - all instruments self-administered:
60
N.A. Bonfils, et al.
Table 2a
Methods and results on quality of life of the randomized-controlled trials.
Studies Design and setting Population Intervention Length Instrument Results on QOL: Results on QOL: 1 st or 2ary Primary outcome: Bias
Significant significant difference outcome Significant difference
difference between between pre- and between groups
groups post-treatment (Yes/No)
(Yes/No) (Yes/No)
Oei et al. Randomized 55 pathological Self-help cognitive behavioral 7 weeks WHOQOL- Not studied Yes 2 Significant difference SB: Unclear, insufficient
(2018) controlled trial gamblers treatment gamblers BREF between pre- and post- information about the
Australia treatment sequence generation
process
PB: Yes, no blinding
DB: Yes, no blinding
AB: No
RB: No
McIntosh et al. Randomized 77 pathological Either manualised Cognitive Behavior 6 months SF-12 Not studied Mental subscale: Yes 2 Composite outcome SB: Unclear, insufficient
(2016) controlled trial gamblers Therapy (CBT1) then MBI Physical sub-scale: measure: no information about the
Australia (Mindfulness-Based Intervention)a No significant difference sequence generation
61
either MBI then CBT1a between the 3 groups. process
either CBT-based treatment as usual Significant difference PB: Yes, no blinding
between pre- and post- DB: Yes, no blinding
treatment AB: No
RB: No
Carlbring and Simple-blind, wait- 66 pathological A wait-list control was compared with 36 months QOLI Not studied Yes 2 NODSb: significant SB: No
Smit list controlled gamblers an 8-week Internet-based cognitive difference between PB: Yes, no blinding
(2008) randomized trial behavior therapy program. pre- and post- DB: No
Sweden Comparison between groups were not treatment AB: No
studied because people on the waiting RB: No
list, for ethical reasons, received
treatment before the follow-up data
were collected
SB: selection bias; PB: performance bias; DB: detection bias; AB: attrition bias; RB: reporting bias.
QOL : Quality of life.
a
2 weeks Wash out period between the 2 interventions.
b
NODS : National Opinion Research Center DSM Screen for Gambling Problems.
Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
N.A. Bonfils, et al.
Table 2b
Methods and results on quality of life of the non-comparative trials.
Studies Design and setting Population Intervention Length of Instrument Results on QOL: 1 st or 2ary Primary outcome:
the study Significant difference outcome Significant difference
between pre- and post- between pre- and post-
treatment treatment
(Yes/No) (Yes/No)
Garcia-Caballero A non-comparative, single 18 pathological gamblers Motivational intervention and CBT therapy 6 months Q-LES-Q No 2 Yes
et al. (2018) group study
Spain
Kim et al. (2016) An uncontrolled trial 150 pathological gamblers Standard helpline care 12 months EUROHIS-QOL Yes 2 Yes
New- Zealand 8-item index
Manning et al. (2014) A non-comparative, single 284 pathological gamblers A CBT-based therapy 12 weeks PWI Yes 1 Composite outcome
group study measure: Yes
Singapore
Carlbring et al. A non-comparative, single 284 pathological gamblers An 8-week Internet-Based Treatment program 36 months QOLI Yes 1 Composite outcome
62
(2012) group study measure: Yes
Sweden
Lahti et al. (2010) A non-comparative, single 39 pathological gamblers 50 mg naltrexone one hour before gambling or 19 weeks EQ-5D Yes 2 PG-YBOCSb: Yes
group study feeling urges to gamble associated with written
Finland instructions for practicing alternative behaviors.
Grant and Potenza A non comparative and open- 13 subjects with both 1/1-week placebo followed by 11 weeks of 22 weeks QOLI Yes 2 PG-YBOCSb: Yes
(2006) label study with double-blind current pathological escitalopram for non placebo-responders
discontinuation gambling and anxiety subjects
USA 2/ Subjects who responded were offered
inclusion in a double-bind discontinuation phase
(2 wash-out weeks for those randomized to
placebo)
Zimmerman et al. A non comparative and open- 15 pathological gamblers Citalopram 12 weeks Q-LES-Q Yes 2 Composite outcome
(2002) label study measure: Yes
USA
a : CBT : cognitive-behavioral therapy ; b : PG-YBOCS : Yale-Brown Obsessive Compulsive Scale Modified for Pathological Gambling.
QOL : Quality of life.
Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
Table 2c
Methods and results on quality of life of the epidemiological studies.
Studies Design / Objective Population Main outcome(s) measure(s) Instruments Main Result(s) on QOL Main Result(s)
Setting
N.A. Bonfils, et al.
Ekholm et al. Cross- To investigate the association 14 924 individuals with current, SF-12 SF-12 Both current and previous problem gambling were negatively
(2018) sectional between problem gambling and past and never problem gambling associated with physical and mental-health problems and QOL.
study health-related quality of life, stress,
Denmark pain or discomfort and the use of
analgesics and sleeping pills
Harries et al. Cross- To examine the differences in three 94 pathological gamblers who did Hamilton and depression anxiety scale QOLI Pathological gamblers seeking Pathological gamblers seeking
(2018) sectional groups of pathological gamblers. not seek treatment; 106 Yale brown obsessive compulsive scale therapy or medical treatments therapy or medical treatments
study pathological gamblers who sought for pathological gambling (PG-YBOCS) had the worst QOL. had lost more money in the past
USA therapy ; 680 pathological Amount lost to gambling past year year due to gambling and scored
gamblers who sought medication higher on the PG-YBOCS
therapy
Black et al. Cross- To compare evolution between 53 older adults with pathological Gambling levels of Shaffer and Hall SF-36 No reported result No difference between the
(2017) sectional older adults and younger adults gambling (≥ 60 years), and two groups.
study comparison groups including 72
USA younger
adults with pathological gambling
(< 40 years) and 50 older adults
without pathological gambling (≥
60 years)
Blum et al. Cross- To characterize neurocognitive 26 problem gamblers (defined as Gambling behavior assessment QOLI Lower QOL among problem Problem gambling with antisocial
(2017) sectional dysfunction in problem gamblers those meeting ≥1 DSM-5 Psychiatric disorders assessment gambling with antisocial personality disorder was
study with co-occurring antisocial diagnostic criteria for gambling Impulsivity assessment personality disorder than among associated with significantly
USA personality disorder disorder) with a lifetime history of Cognitive assessment controls elevated gambling disorder
antisocial personality disorder and symptoms, greater psychiatric
63
266 controls comorbidity, higher impulsivity
questionnaire scores on the
Barratt Impulsiveness Scale
(d = 0.4) and Eysenck
Impulsivity Questionnaire
(d = 0.5), and impaired cognitive
flexibility (d = 0.4), executive
planning (d = 0.4),
Chamberlain Cross- To identify subtypes of gambling in 582 non-treatment seeking young Gambling behavior criteria QOLI QOL was impaired in problem Three subtypes of gambling were
et al. (2017) sectional young adults, using latent class adults were recruited from two US - Structured Clinical Interview for and pathological gamblers identified, termed recreational
study analysis, based on individual cities, on the basis of gambling five Gambling Disorder versus recreational gamblers, gamblers (60.2% of the sample;
USA responses from the Structured or more times per year. - Amount lost to gambling past year and problem gamblers did not reference group), problem
Clinical Interview for Gambling differ significantly from gamblers (29.2%), and
Disorder pathological gamblers on this pathological gamblers (10.5%).
measure.
Medeiros et al. Cross- To investigate clinical and 448 adults diagnosed with Duration of illness QOLI Negative correlation between Negative correlation between
(2017) sectional neurocognitive characteristics gambling disorder. Lag between first gambling and duration of illness and QOL duration of illness and lag
study associated with different duration onset of gambling disorder between first gambling and onset
USA of illness Gambling frequency of gambling disorder
Lifetime prevalence of co occuring Lifetime history of alcohol use
psychiatric disorders disorder associated with a longer
Family history of alcohol use disroder duration of gambling disorder
or gambling disorder Presence of a first-degree relative
with history of alcohol use
disorder associated with a more
extended course of gambling
disorder
Chamberlain Cross- Structured Clinical Interview for QOLI The gambling disorder group The gambling disorder group had
et al. (2016) sectional Pathological Gambling had a low QOL (mean total score a mean PG-YBOCS score of 24.2
(continued on next page)
Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
Table 2c (continued)
Studies Design / Objective Population Main outcome(s) measure(s) Instruments Main Result(s) on QOL Main Result(s)
Setting
N.A. Bonfils, et al.
study To compare white matter integrity 16 participants with treatment- PG-YBOCS of 29.8, s.d. = 12.6) according (s.d. = 4.7) and a mean G-SAS
USA in patients with gambling disorder resistant gambling disorder and 15 Gambling Symptom Assessment Scale to the QOLI but relatively low score of 33.8 (s.d. = 7.1) both of
with healthy controls healthy controls depression scores on the HRSD which corresponded to a severe
(mean total score 8.3, gambling disorder
s.d. = 4.5).
No measure of QOLI in control
group
Loo et al. (2016) Cross- To provide an evaluation of the 445 Macao and Australian young Gambling Urge Scale WHOQOL-BREF Pathological gamblers reported 1/ Alcohol and erroneous
sectional predictive ability of psychological adults Gambling Related Cognitions Scale significantly lower QOL in all cognitions increase pathological
study variables WHOQOL-BREF domains. gambling behavior.
China - The Alcohol Use Disorders 2/ Erroneous gambling-related
& Australia Identification Test cognitions serve as a mediator for
the predictive relationship
between gambling urge and
pathological gambling.
Medeiros et al. Cross- To evaluate the association 143 non-treatment seeking young Severity of anxiety symptoms QOLI The severity of anxiety Positive correlation between
(2016) sectional between anxiety symptoms, adults (aged 18–29 years), in variables symptoms was negatively anxiety severity and gambling
study gambling activity, and which 63 individuals (44.1%) were Clinical variables correlated with the quality of severity measured by the number
USA neurocognition across the classified as recreational gamblers, Neurocognitive variables life of DSM- 5 gambling disorder
spectrum of gambling behavior 47 (32.9%) as having criteria met
subsyndromal GD, and 33 (23.1%)
met criteria for GD.
Manning et al. Cross- To examine rates and predictors of 2187 Singaporean patients with Rates of suicidal behaviors PWI Among all addiction patients, 1/ Rates of suicidal ideation
(2015) sectional suicide behaviors among substance drug (n = 879), alcohol (n = 754), significantly lower QOL among significantly higher among
study use disorder and gambling disorder or gambling (n = 554) disorders those with suicidal thoughts/ gambling than substance use
64
Singapore patients entering an outpatient treatment plan and among those reporting patients.
service lifetime attempts 2/Comordid disorders : 32.5%,
38% and 40% among those
reporting thoughts, plan and
lifetime attempts.
3/ Comorbidity, debt, gender and
being a gambler are significant
predictors of suicidal behaviors
Oei and Raylu Cross- To assess whether cognitive and 139 pathological gamblers Correlation between cognitive/ WHOQOL-BREF QOL significantly correlated Perceived gambling refusal self-
(2015) sectional psychosocial variables can predict psychosocial variables with PG with pathological gambling efficacy, cognitions of inability to
study gambling behavior in a clinical Prediction of pathological gambling stop gambling, hazardous
Australia sample scores (SOGS CGPI) drinking behaviors and poor QOL
related to environment
significantly predicted
pathological gambling.
Subramian et al. Cross- To examine the co morbidity of Singapore residents aged 18 years Prevalence of pathological gambling EQ-5D Significantly lower QOL among 1/Pathological gambling
(2015) sectional pathological gambling with other and above, randomly selected from and odds ratio for lifetime DSM-IV EQ-VAS pathological gamblers than non- associated significantly with
study mental and physical disorders as a database obtained from a mental disorders and chronic physical gamblers/non-problem major depressive disorder, bipolar
Singapore well as to examine health related national registry. conditions among non-gamblers/non- gamblers disorder, alcohol abuse/
QOL perceived by those with 5986 subjectsB problem gamblers and pathological dependence, and any co morbid
pathological gambling gamblers. mental disorder including
nicotine dependence.
2/ Pathological gambling
associated significantly with
hypertension, ulcer and chronic
inflammatory bowel disease and
any co morbid chronic physical
condition.
Kohler (2014) Quality adjusted life year
(continued on next page)
Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
Table 2c (continued)
Studies Design / Objective Population Main outcome(s) measure(s) Instruments Main Result(s) on QOL Main Result(s)
Setting
N.A. Bonfils, et al.
Cross- To estimate the Health Related 52 pathological gamblers recruited SF-6D Significantly lower QOL among Pathological gambling
sectional Quality of life costs of Gambling from treatments centers in EQ-VAS the pathological gamblers than significantly associated with a
study addiction Switzerland and 93 non-gamblers among the control group decrease in the quality of life by
Switzerland subjects (representative of the 0.076 quality adjusted life year.
Swiss population)
Suomi et al Cross- To identify subtypes of gamblers 212 new clients in gambling Subtype of gamblers WHOQOL-BREF Significantly lower quality of 4 identified subgroups of
(2014) sectional and determine if the subgroups are programs life among multimorbid participants on the basis of
study different on demographic factors, gamblers than among the others psychological distress, alcohol
Australia gambling, and general well being abuse and impulsivity: gamblers
factors. with co morbid psychological
problems (35%), pure gamblers
(27%), gamblers with co morbid
alcohol abuse (25%), mutimorbid
gamblers (13%).
Black et al. Cross- To assess self-reported chronic 95 subjects with pathological Prevalence and adjusted odds ratio SF-36 Significantly lower QOL among Pathological gambling associated
(2013) sectional medical conditions, medication gambling between pathological gambling and pathological gamblers on all but with:
study. usage, lifestyle choices, health care 91 controls medical health, lifestyle choices, one SF-36 subscale (physical - Obesity
USA utilization, quality of life variables, medication usage and health care role subscale). - Chronic medical conditions
and body mass index in persons utilization - Poor lifestyle choices
with and without pathological - Use of costly forms of medical
gambling care
Manning et al. Cross- To examine quality of life in 3 260 alcohol-dependent, 282 drug- QOL PWI 1/ Significantly lower QOL among pathological gamblers than among
(2012) sectional different addiction populations at dependent and 132 pathological alcohol patients. But no significantly differences in personal
study their first visit to treatment gambling outpatients relationships item
Singapore 2/ Alcohol patients scored significantly higher on the PWI compared
65
with drug patients
Chou and Afifi Cohort study To assess the association of past- 33 231 subjects from nationally Odds ratios for the relation between SF-12 Reduced health-related QOL among disordered gamblers
(2011) USA year disordered gambling with the representative US sample past-year disordered gambling and the
incidence of Axis I psychiatric interviewed in 2000-2001 and incidence of Axis I psychiatric
disorders at follow-up 3 years later, 2004-2005 disorders during the 3-year follow-up
after adjusting for demographic period
variables, medical conditions,
health-related quality of life, and
stressful life events
Mythily et al. Cross- To examine the quality of life in 40 subjects with compulsive QOL WHOQOL-BREF Significantly lower QOL among pathological gamblers than controls
(2011) sectional pathological gamblers in a gambling behavior (SOGS > 5)
study multiracial population and 40 controls
Singapore
Kennedy et al. Cross- To investigate the frequency of 579 participants recruited from 5 Frequency of gambling Q-LES-Q Lower QOL among problem Prevalence of problem gambling
(2010) sectional gambling in people who have been sites in Canada and 1 in the USA gamblers compared with non- not differs significantly between
study diagnosed with major depressive and meeting criteria for lifetime problem gamblers both for the the MDD and the BD groups.
Canada /USA disorder (MDD) or bipolar disorder MDD or bipolar disorder I or II MDD and bipolar disorder
groups
Najavits et al. Cross- To compare pathological 106 adults recruited from the Mean differences between the 3 Q-LES-Q Q-LES-Q : High rates of additional co-
(2010) sectional gambling, posttraumatic stress community: 35 with current groups, for sociodemographic, & 1/Significant differences on the occuring disorders and suicidality
study disorder (PTSD) and their co- pathological gamblers, 36 with psychopathology, functioning, SF-12 overall score and 3 of the 5 in PTSD and BOTH
USA occurrence current PTSD, 35 with both cognition and life history variables subscales (non significant on the 59% and 34% reported
2 subscales: physical and respectively a family history of
satisfaction with medication) substance use disorder and
between the 3 groups. gambling problems
2/Pathological gamblers
reported higher QOL than PTSD
and BOTH
SF-12: pathological gamblers
(continued on next page)
Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
Table 2c (continued)
Studies Design / Objective Population Main outcome(s) measure(s) Instruments Main Result(s) on QOL Main Result(s)
Setting
N.A. Bonfils, et al.
66
Scherrer et al. Cross- To model differences in the health- Male twin members of a Vietnam QOL SF-8 1/ No significantly differences between groups in the physical and
(2005) sectional related quality of life among non- registry: 53 pathological gamblers, bodily pain subscales after adjusting for confounding factors
Vietnam problem gamblers, problem 270 subclinical problem gamblers, 2/ For the mental subscale, PG have lower score than problem gamblers
gamblers, and pathological and 1346 non-problem gamblers and those have lower score than non-problem gamblers
gamblers 3/ After adjustment for covariates, no significantly difference between
non-problem gambling index twins and affected co-twins
Black et al. Cross- To assess health-related quality of 30 pathological gamblers QOL SF-36 In comparison with US population norm, lower QOL among
(2003) sectional life and family psychiatric history pathological gamblers in several import domains, including mental
study in PG health, physical functioning, bodily pain, and general health.
USA
3.1. Description of the instruments and their psychometric properties between the different instruments on the assessed areas. Most of the
instruments assessed social relationships (8/9) whereas EQ-5D did not.
The instruments differed widely in number of domains and items, Moreover, only QOLI assessed “problems with children” and only
scaling, scoring and psychometric properties. The complete description WHOQOL-BREF did not assess family relationships. Social activities
of these instruments is provided in Table 3. form the most explored domain across the instruments. Finally, only 3
The psychometric properties of the included scales have been poorly instruments out of 6 assess the financial concerns (QOLI, WHOQOL-
studied in the gambling population. We report those given in the in- BREF and Q-LES-Q).
cluded articles. The SF-12, SF-8 and SF-6D derive from the SF-36. In a
sample of 77 pathological gamblers, internal consistency (Cronbach’s
alpha) of the SF-12 domains ranged from 0.76 to 0.77, indicating rather 3.3. Meta-analysis: effect size of QOL instruments
good reliability (McIntosh et al., 2016). No more psychometric prop-
erties have been reported in the gambling population for SF scales. In a We studied the effect size of the QOL/HRQOL instruments and
sample of 139 individuals diagnosed with gambling disordered, internal gambling outcomes after an intervention. To this end, we only selected
consistency of the WHOQOL-BREF domains measured ranged from 0.66 trials and performed a meta-analysis on gambling outcomes and QOL
to 0.84, indicating rather good reliability (Oei and Raylu, 2015). In a outcomes. Three trials were excluded because of a lack of documenta-
sample of 260 alcohol-dependent, 282 drug-dependent, and 132 pa- tion on QOL or outcome results. When combining the seven trials, our
thological gambling outpatients, internal consistency of the PWI was meta-analysis showed these results: the global effect size for QOL/
0.87 (Manning et al., 2012). To the best of our knowledge, psycho- HRQOL was 0.54 (95%CI 0.34-0.73); concerning gambling outcome,
metric properties of EQ-5D, QOLI and Q-LES-Q have not been reported the global effect size was 1.550 (95%CI 1.20-1.90) (Fig. 2). Thus the
in the gambling-disordered population. measurement of QOL with non-gambling specific instruments showed a
significant medium-range effect size, significantly smaller than the
large-range gambling outcome effect size. The heterogeneity was quite
3.2. Analysis of the current concept of quality of life in gambling-disordered high, rather similar for both outcomes: I2 = 8782 for gambling outcome
subjects and I2 = 8320 for QOL/HRQOL. A visual inspection of the funnel plot
for open-label studies revealed asymmetry, indicating possible pub-
The twenty-six domains investigated by the QOL/HRQOL instru- lication bias (Fig. 3). Using Trim and Fill the imputed point estimate is
ments of the included articles are presented in Table 4. We categorized barely modified: 0.54 (95%CI 0.34-0.73) for the QOL/HRQOL outcome
these domains into the following broader categories: relationships with and 1.53 (95%CI 1.17–1.89) for the gambling outcome.
others, activities, physical state, psychological state, financial concerns,
medical care and satisfaction with life. Some disparities can be noticed
67
N.A. Bonfils, et al. Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
Table 3
Description of the used instruments.
Instruments Domains Domain description Scaling & scoring
(items)
* SF-36 : The Medical Outcomes Study Item Short-Form Health Survey ; EQ-5D : The European Quality of Life Questionnaire; WHOQOL-BREF: The World Health
Organization Quality of Life Assessment Instrument ; Q-LES-Q: The Quality of Life Enjoyment and Satisfaction Questionnaire ; QOLI : The Quality of Life Inventory ;
PWI: The Personal Well-Being.
Table 4
Domains investigated by quality of life instruments used in included studies.
Categories Domains SF-36 QOLI WHOQOL-BREF / EURHOHIS-QOL 8-item index Q-LES-Q EQ-5D PWI
SF-12 EQ-VAS
SF6-D/SF-8
68
N.A. Bonfils, et al. Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
69
N.A. Bonfils, et al. Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
in seven broader categories: relationships with others, activities, phy- Finally, QOL instruments should be designed to reflect the impact of
sical state, psychological state, financial concerns, medical care and diseases and interventions in daily life from the subject’s perspective
satisfaction with life. Nevertheless, these instruments are general (i.e. (Leplège, 1999). However, most items of these instruments have been
unspecific) QOL/HRQOL instruments, and might not evaluate the defined by experts to answer their own questions. Therefore a tension
whole range of impacted QOL/HRQOL caused by gambling disorders. may appear between the objective of the experts and the impact of the
Only a few studies revealed the results of the scale sub-scores, while or treatment as perceived by subjects. Despite being self-administered
others gave an overall result. When we looked at the studies giving the and requesting subjects’ subjective answers, the instruments are not
results of the sub scores, only one study found a significant result on all necessarily a good reflection of subjects concerns. Moreover, if these
the sub-scores including mental health, physical functioning, bodily general instruments can partially explore the occurrence of adverse
pain, and general health (Black et al., 2003). The other studies found consequences related to gambling disorder, it is important to evaluate
mental sub-score significantly associated with gambling but no sig- the subjective value of these consequences for each subject. It is clear
nificant results regarding physical, bodily pain, vitality or general that two subjects can assign different weights to a single negative
health sub-scores (Black et al., 2013; Mason and Arnold, 2007; consequence related to his gambling addiction. While many studied
McIntosh et al., 2016; Morasco and Petry, 2006; Najavits et al., 2010; areas are common across the six instruments, the wording of the
Scherrer et al., 2005). Previous studies focused on harms caused by question and thus the answers can differ widely. For instance, the
gambling (Commission, 2010; Langham et al., 2015) or on HRQOL wording of the question on the financial aspect varies between two
areas impacted by gambling (Bonfils et al., 2019). Bonfils et al. (2019) questionnaires. The Q-LES-Q is interested in the satisfaction with the
aimed to describe HRQOL impacted by problem gambling from a pa- economic status during the past week. The WHOQOL-BREF asks if the
tients’ perspective, through a qualitative study. Financial pressure, re- subject had enough money to meet his (her) needs in the last two
lationships deterioration, negative emotions, loneliness, feeling of in- weeks. The different wordings and the different periods explored could
comprehension, preoccupation with gambling and avoidance of helping induce dissimilar answers.
relationship were found to be HRQOL domains impacted by gambling The Food and Drug Administration (FDA) encourages specific
(Bonfils et al., 2019). Loneliness, feeling of incomprehension, pre- Patients Reported Outcomes (PROs) instrument to support labeling
occupation with gambling and avoidance of helping relationship were claims of medical products (FDA, 2009). PRO instruments insure the
not mentioned in HRQOL instruments previously used in problem subjectivity of the outcome, but not its clinical pertinence. A pertinent
gambling (Bonfils et al., 2019). Moreover, Luquiens et al. (Luquiens PRO instrument should be patient-focused. One way to reach this goal
et al., 2015) studied the impact of alcohol use disorder on HRQOL would be to develop the instrument from the subjects themselves, via
through a qualitative analysis. To this end, they conducted focus groups qualitative analysis of interviews with gambling-disorder subjects. This
with 38 alcohol use disorder subjects and found seven areas of impact: methodology would allow identification of the most impacted domains
relationships, emotional impact, living conditions, control, activities, of life in this particular population and those aspects that are most
looking after self and sleep . Only the first three areas were common for sensitive to change, thereby promising real usefulness in assessing the
both addictive disorders. Although substance use disorder and gam- efficacy of interventions among gambling disorder subjects. The de-
bling disorder share common characteristics, people with gambling velopment of a patient-focused rigorous PRO QOL instrument specific
disorders also have specific characteristics (Grant and Chamberlain, to subjects with gambling disorders would allow clinicians to have a
2015). It seems that general QOL/HRQOL instruments do not capture relevant instrument to administer to their patients (Arpinelli and Bamfi,
gambling specificities. It was for this reason that recent Burden of 2006).
Disease assessments of gambling in New Zealand and Australia utilized There are several limitations to our literature review. First of all, we
time trade off and visual analogic scale elicitation methods to direct did not conduct an in-depth search of the grey literature. However, we
assess community and expert opinions, rather inferring QOL from looked at the grey literature with the same keywords on the database
generic instruments (Browne et al., 2017). The six different instruments OpenSIGLE, recommended by Cochrane, and did not find any results.
listed in the included articles are very heterogeneous regarding the Secondly, our review was limited to English-language articles. Some
QOL/HRQOL domains they assess (Table 3). It would be interesting to studies may not have been included due to this. With regard to meta-
standardize QOL/HRQOL instruments in this population in order to analysis, we can also highlight some limitations. First of all, the in-
allow comparisons across different studies. However, there is currently cluded articles had a great heterogeneity of judgment criteria. Finally,
no specific instrument to assess a change in QOL/HRQOL in subjects we did not take into account the biases of the trials for inclusion in the
with a gambling disorder while there are instruments for specific dis- meta-analysis.
eases in many specialties (such as oncology (Macquart-Moulin et al.,
2000) or neurology (Fujishima-Hachiya and Inoue, 2012; Tan et al.,
2005) or gastroenterology (Zingone et al., 2013)), and in the addiction 5. Conclusion
field in particular (as alcoholism (Luquiens et al., 2016, 2015; Neale
et al., 2016)). A successful intervention can be indirectly measured by consequent
However, it seems interesting to look at which general instruments increases in the QOL. The currently used general instruments are effi-
of QOL/HRQOL seem most appropriate to use in this field of research, cient to measure a significant change after an intervention but might
taking into account psychometric properties and content. SF, WHOQOL, not evaluate specific areas of health related QOL impacted by gambling
PWI are the only instruments that are partially psychometrically vali- disorders. Moreover, they are not totally validated in the problem
dated with a correct internal consistency in gambling population. Only gambling population and are probably less sensitive than gambling
SF scales are HRQOL instruments; WHOQOL and PWI are QOL instru- behavioral criteria. Despite being self-administered and requesting
ments. Nevertheless, SF scales fail to capture functioning in do- subjective answers, these instruments do not necessarily explore the
mains—especially financial concerns or self-esteem —that are im- entire spectrum of patients’ concerns on the impact on QOL of gambling
portant to gambling disorder-affected populations. The same is true for disorder and do not take into account the patient’s subjectivity. Patients
PWI, which also does not explore these domains or psychological well- should contribute to define the content of impact on QoL of gambling
being. The scale with both psychometric qualities and the most ap- disorder. Such an approach could serve as the basis for the future de-
propriate content for gambling is probably WHOQOL which offers a velopment of a tool specifically designed for gambling disorders in
very promising alternative, yielding scores in psychological and social order to assess health-related QoL.
functioning; living environment; and an overall satisfaction score
(Laudet, 2011).
70
N.A. Bonfils, et al. Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
71
N.A. Bonfils, et al. Neuroscience and Biobehavioral Reviews 104 (2019) 58–72
recombinant granulocyte colony-stimulating factor and repeated stem-cell support Stud. 34, 581–595. https://doi.org/10.1007/s10899-017-9723-1.
for inflammatory breast cancer patients: does impact on quality of life jeopardize Pickering, D., Keen, B., Entwistle, G., Blaszczynski, A., 2017. Measuring treatment out-
feasibility and acceptability of treatment? J. Clin. Oncol. Off. J. Am. Soc. Clin. Oncol. comes in gambling disorders: a systematic review. Addict. Abingdon Engl. https://
18, 754–764. https://doi.org/10.1200/JCO.2000.18.4.754. doi.org/10.1111/add.13968.
Manning, V., Gomez, B., Guo, S., Low, Y.D., Koh, P.K., Wong, K.E., 2012. An exploration Rosenthal, R., 1991. Meta-Analytic Procedures for Social Research. Rosenthal.
of quality of life and its predictors in patients with addictive disorders: gambling, Scherrer, J.F., Xian, H., Shah, K.R., Volberg, R., Slutske, W., Eisen, S.A., 2005. Effect of
alcohol and drugs. Int. J. Ment. Health Addict. 10, 551–562. https://doi.org/10. genes, environment, and lifetime co-occurring disorders on health-related quality of
1007/s11469-011-9349-y. life in problem and pathological gamblers. Arch. Gen. Psychiatry 62, 677–683.
Manning, V., Koh, P.K., Yang, Y., Ng, A., Guo, S., Kandasami, G., Wong, K.E., 2015. https://doi.org/10.1001/archpsyc.62.6.677.
Suicidal ideation and lifetime attempts in substance and gambling disorders. Schmidt, S., Mühlan, H., Power, M., 2006. The EUROHIS-QOL 8-item index: psychometric
Psychiatry Res. 225, 706–709. https://doi.org/10.1016/j.psychres.2014.11.011. results of a cross-cultural field study. Eur. J. Public Health 16, 420–428. https://doi.
Manning, V., Ng, A., Koh, P.K., Guo, S., Gomathinayagam, K., Wong, K.E., 2014. org/10.1093/eurpub/cki155.
Pathological gamblers in Singapore: treatment response at 3 months. J. Addict. Med. Scott, J., Huskisson, E.C., 1976. Graphic representation of pain. Pain 2, 175–184.
8, 462–469. https://doi.org/10.1097/ADM.0000000000000082. Slutske, W.S., Jackson, K.M., Sher, K.J., 2003. The natural history of problem gambling
Mason, K., Arnold, R., 2007. Problem gambling risk factors and associated behaviours and from age 18 to 29. J. Abnorm. Psychol. 112, 263–274.
health status: results from the 2002/03 New Zealand Health Survey. N. Z. Med. J. Subramaniam, M., Abdin, E., Vaingankar, J.A., Wong, K.E., Chong, S.A., 2015. Comorbid
120, 9–20. physical and mental illnesses among pathological gamblers: results from a population
McHorney, C.A., Ware, J.E., Raczek, A.E., 1993. The MOS 36-Item Short-Form Health based study in Singapore. Psychiatry Res. 227, 198–205. https://doi.org/10.1016/j.
Survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical psychres.2015.03.033.
and mental health constructs. Med. Care 31, 247–263. Suomi, A., Dowling, N.A., Jackson, A.C., 2014. Problem gambling subtypes based on
McIntosh, C.C., Crino, R.D., O’Neill, K., 2016. Treating problem gambling samples with psychological distress, alcohol abuse and impulsivity. Addict. Behav. 39, 1741–1745.
cognitive behavioural therapy and mindfulness-based interventions: a clinical trial. J. https://doi.org/10.1016/j.addbeh.2014.07.023.
Gambl. Stud. Co-Spons. Natl. Counc. Probl. Gambl. Inst. Study Gambl. Commer. Szabo, S., WHOQOL Group, 1996. The World Health Organization Quality of Life
Gaming. https://doi.org/10.1007/s10899-016-9602-1. (WHOQOL) Assessment Instrument.
Medeiros, G.C., Redden, S.A., Chamberlain, S.R., Grant, J.E., 2017. Gambling disorder: Tan, E.K., Fook-Chong, S., Lum, S.-Y., Thumboo, J., 2005. Validation of a short disease
association between duration of illness, clinical, and neurocognitive variables. J. specific quality of life scale for hemifacial spasm: correlation with SF-36. J. Neurol.
Behav. Addict. 6, 194–202. https://doi.org/10.1556/2006.6.2017.029. Neurosurg. Psychiatry 76, 1707–1710. https://doi.org/10.1136/jnnp.2005.065656.
Medeiros, G.C., Sampaio, D.G., Leppink, E.W., Chamberlain, S.R., Grant, J.E., 2016. Walker, M., Toneatto, T., Potenza, M.N., Petry, N., Ladouceur, R., Hodgins, D.C., el-
Anxiety, gambling activity, and neurocognition: a dimensional approach to a non- Guebaly, N., Echeburua, E., Blaszczynski, A., 2006. A framework for reporting out-
treatment-Seeking sample. J. Behav. Addict. 5, 261–270. https://doi.org/10.1556/ comes in problem gambling treatment research: the Banff, Alberta Consensus.
2006.5.2016.044. Addiction 101, 504–511. https://doi.org/10.1111/j.1360-0443.2005.01341.x.
Morasco, B.J., Petry, N.M., 2006. Gambling problems and health functioning in in- Ware, J., Kosinski, M., Keller, S.D., 1996. A 12-Item Short-Form Health Survey: con-
dividuals receiving disability. Disabil. Rehabil. 28, 619–623. https://doi.org/10. struction of scales and preliminary tests of reliability and validity. Med. Care 34,
1080/09638280500242507. 220–233.
Mythily, S., Edimansyah, A., Qiu, S., Munidasa, W., 2011. Quality of life in pathological Ware, J.E., 2000. SF-36 health survey update. Spine 25, 3130–3139.
gamblers in a multiethnic Asian setting. Ann. Acad. Med. Singapore 40, 264–268. Ware, J.E., GlaxoSmithKline, 2001. How to Score and Interpret Single-item Health Status
Najavits, L.M., Meyer, T., Johnson, K.M., Korn, D., 2010. Pathological gambling and Measures: a Manual for Users of the of the SF-8 Health Survey : (with a Supplement
posttraumatic stress disorder: a study of the Co-morbidity versus each alone. J. on the SF-6 Health Survey). QualityMetric, Inc. ; Health Assessment Lab, Lincoln, RI,
Gambl. Stud. 27, 663–683. https://doi.org/10.1007/s10899-010-9230-0. Boston, MA.
Neale, J., Vitoratou, S., Finch, E., Lennon, P., Mitcheson, L., Panebianco, D., Rose, D., Wu, J.Q., Appleman, E.R., Salazar, R.D., Ong, J.C., 2015. Cognitive behavioral therapy for
Strang, J., Wykes, T., Marsden, J., 2016. Development and validation of “SURE”: a insomnia comorbid with psychiatric and medical conditions: a meta-analysis. JAMA
patient reported outcome measure (PROM) for recovery from drug and alcohol de- Intern. Med. 175, 1461–1472. https://doi.org/10.1001/jamainternmed.2015.3006.
pendence. Drug Alcohol Depend. 165, 159–167. https://doi.org/10.1016/j. Zimmerman, M., Breen, R.B., Posternak, M.A., 2002. An open-label study of citalopram in
drugalcdep.2016.06.006. the treatment of pathological gambling. J. Clin. Psychiatry 63, 44–48.
Oei, T.P.S., Raylu, N., 2015. Cognitive and psychosocial variables predicting gambling Zingone, F., Iavarone, A., Tortora, R., Imperatore, N., Pellegrini, L., Russo, T., Dorn, S.D.,
behavior in a clinical sample. Int. J. Ment. Health Addict. 13, 520–535. https://doi. Ciacci, C., 2013. The Italian translation of the celiac disease-specific quality of life
org/10.1007/s11469-015-9555-0. scale in celiac patients on gluten free diet. Dig. Liver Dis. Off. J. Ital. Soc.
Oei, T.P.S., Raylu, N., Lai, W.W., 2018. Effectiveness of a self help cognitive behavioural Gastroenterol. Ital. Assoc. Study Liver 45, 115–118. https://doi.org/10.1016/j.dld.
treatment program for problem gamblers: a randomised controlled trial. J. Gambl. 2012.10.018.
72