Psychology of Addictive Behaviors 2014, Vol. 28, No. 4, 1190 –1197

© 2014 American Psychological Association 0893-164X/14/$12.00 http://dx.doi.org/10.1037/a0037603

Gambling Disorder: Estimated Prevalence Rates and Risk Factors in Macao Anise M. S. Wu

Mark H. C. Lai

University of Macau

Texas A&M University

Kwok-Kit Tong

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

University of Macau An excessive, problematic gambling pattern has been regarded as a mental disorder in the Diagnostic and Statistical Manual for Mental Disorders (DSM) for more than 3 decades (American Psychiatric Association [APA], 1980). In this study, its latest prevalence in Macao (one of very few cities with legalized gambling in China and the Far East) was estimated with 2 major changes in the diagnostic criteria, suggested by the 5th edition of DSM (APA, 2013): (a) removing the “Illegal Act” criterion, and (b) lowering the threshold for diagnosis. A random, representative sample of 1,018 Macao residents was surveyed with a phone poll design in January 2013. After the 2 changes were adopted, the present study showed that the estimated prevalence rate of gambling disorder was 2.1% of the Macao adult population. Moreover, the present findings also provided empirical support to the application of these 2 recommended changes when assessing symptoms of gambling disorder among Chinese community adults. Personal risk factors of gambling disorder, namely being male, having low education, a preference for casino gambling, as well as high materialism, were identified. Keywords: gambling, disorder, diagnostic criteria, prevalence, Chinese

Psychiatric Association [APA], 1980). This kind of behavioral addiction was previously labeled as pathological gambling, but was recently renamed as gambling disorder under the category of substance-related and addictive disorders in the latest edition of DSM (DSM-5; APA, 2013). Disordered gamblers often experience similar symptoms to those of alcoholics and drug addicts, and report strong craving. The diagnostic criteria of gambling disorder also mimic those of substance dependence and include preoccupation with gambling, tolerance development (increasing wager amount), withdrawal symptoms like restlessness and irritability, and jeopardizing significant relationships, employment, and so forth because of gambling. One should note that there are two major changes in the criteria and threshold for diagnosing this mental disorder in DSM-5. First, the criterion “has committed illegal acts such as forgery, fraud theft or embezzlement to finance gambling” has been removed and therefore only nine criteria are listed in DSM-5 for gambling disorder. This criterion is believed to contribute little to diagnostic accuracy of the disorder (Petry et al., 2014). For example, in a study of a group of outpatients in a French addiction clinic, the DSM score calculated with or without the “Illegal Acts” criterion yielded no practical differences in estimated prevalence, classification sensitivity and specificity, Cronbach’s alpha, and correlation with gambling severity (Denis, Fatséas, & Auriacombe, 2012). The second major change is the reduction of the threshold of diagnosis from five to four criteria. The lowered threshold is believed to improve diagnostic accuracy, with some empirical support in treatment-seeking samples in the United States, Spain, and France (Denis et al., 2012; Jiménez-Murcia et al., 2009; Stinchfield, 2003). Petry et al. (2014) also briefly reported the findings of their study with a community sample and indicated

The Macao Special Administrative Region of the People’s Republic of China (Macao) is one of the few cities in China and the Far East where gambling is legal. In this city with an area of only around 30 square kilometers and a population of 557,400, there are 34 casinos with more than 5,300 casino tables and 160,000 slot machines, as well as other gaming facilities for horse-racing, greyhound-racing, pacapio lottery, and sports betting (Macao Statistics & Census Service, 2012a). The gaming industry brought more than 33 billion USD of gross revenue in 2011, ranking number 1 in the world. It constitutes two fifths of the GDP and 23.0% of the total employment in Macao. In response to the rapid development of the gaming industry, local residents’ response to the proliferation of legal gambling in Macao should be regularly monitored by the local government, and research concerning both prevalence rate and risk factors of gambling disorder is necessary.

Diagnostic Criteria and Threshold of Gambling Disorder An excessive, problematic gambling pattern has been regarded as a mental disorder in the Diagnostic and Statistical Manual for Mental Disorders (DSM) for more than three decades (American

This article was published Online First August 18, 2014. Anise M. S. Wu, Department of Psychology, University of Macau; Mark H. C. Lai, Department of Educational Psychology, Texas A&M University; Kwok-Kit Tong, Department of Psychology, University of Macau. Correspondence concerning this article should be addressed to Anise M. S. Wu, Department of Psychology, Faculty of Social Sciences, University of Macau, Av. Padre Tomás Pereira, Taipa, Macao, China. E-mail: [email protected] 1190

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

GAMBLING DISORDER IN MACAO

that, similar to the findings among treatment-seeking samples, the prevalence rate, internal consistency, factors structure, and classification accuracy remained more or less the same or even improved if the Illegal Acts criterion was dropped and the cut-off score was set as four. Although there is some empirical evidence showing that the changes in DSM-5 made little impact on the prevalence rate and improved classification accuracy of disordered gamblers, more diverse research should be done to address this issue (Reilly & Smith, 2013). In particular, the application of these changes on non-Western populations remains unknown. Furthermore, given that the DSM criteria have been widely used as an assessment tool of the severity of gambling disorder, its psychometric properties should be further investigated after the removal of the Illegal Acts criterion.

Estimation of the Prevalence of Gambling Disorder in Macao The reported gambling involvement rates among Macao residents vary in the research with telephone survey design. The estimates of the past-year gambling participation in other surveys ranged from 33.0% to 67.9% (Fong & Ng, 2010; Fong & Ozorio, 2005; Wu, Lai, Tong, & Tao, 2013). The variation may result from temporal or methodological differences. Fong and Ng (2010) also estimated the current prevalence rates of gambling disorder in 2003, 2007, and 2010 according to a Chinese version of the 10 criteria listed in DSM–IV for pathological gambling (APA, 1994) with a threshold of five, and the estimates were 1.8%, 2.6%, and 2.8%, respectively, among 15- to 64-year-old Macao residents. However, little is known about how the changes in criteria and threshold for diagnosis of gambling disorder suggested by DSM-5 may influence the prevalence rates estimated in Chinese populations. In the present study, we investigated whether the estimated prevalence of gambling disorder would vary with different criterion number and threshold of diagnosis (i.e., five of 10 or four of nine DSM criteria1) with a representative adult sample recruited in Macao, China in 2013. Given that DSM criteria have been widely used as an assessment tool, we also investigated whether the Chinese version of nine DSM criteria is a psychometrically sound measurement inventory of the severity of gambling disorder in this sample. Its factor structure, internal consistency, and convergent validity were examined. To demonstrate convergent validity, we also computed the correlations between the number of endorsed DSM criteria and other four indicators of gambling disorder, namely gambling intention, frequency, expenditure, and maximum wager size. These indicators were used in a French sample to evaluate the convergent validity of the DSM-5 criteria (Denis et al., 2012). They were also positively correlated with symptoms of gambling disorder in Chinese samples (Tao, Wu, Cheung, & Tong, 2011; Wu et al., 2013; Wu & Tang, 2012; Wu, Tao, Tong, & Cheung, 2012).

Risk Factors of Gambling Disorder in Macao In order to design better public health policies against gambling disorder in Chinese societies that have a gaming industry like Macao, risk factors that contribute to gambling disorder should be identified. Therefore, after evaluating the new assessment criteria

1191

and estimating the problem severity, we took one step further and tested two psychological (i.e., materialism and dissatisfaction of life) and four demographic (i.e., gender, age, education, and casino gambling) risk factors of gambling disorder among adults in Macao in this study.

Materialistic Value Materialism is defined as “a set of centrally held beliefs about the importance of possession in one’s life” (Richins & Dawson, 1992, p. 308), and is a more and more important value in modern Chinese culture. Its importance can be reflected in the increasing consumption of luxury goods in China (World Luxury Association, 2011). Materialism has been shown to have relationships with not only gambling participation but also disordered gambler status in American samples (Netemeyer et al., 1998; Fang & Mowen, 2009). Moreover, gambling motive for monetary gain, which is positively associated with Chinese gamblers’ problems (Tao et al., 2011; Wu et al., 2012), increases with materialism (Fang & Mowen, 2009). In the present study, we hypothesized a positive correlation between materialism and the number of DSM criteria endorsed in our Chinese sample.

Dissatisfaction of Life Different facets of subjective well-being, such as life satisfaction, may exert influence and be influenced adversely by addiction development. Mythily, Edimansyah, Qiu, and Munidasa (2011) reported that disordered gamblers scored significantly lower in all domains of quality of life than their sociodemographically matched controls. Life dissatisfaction may lead to negative emotions and thus drive Chinese gamblers to gamble for the sake of boredom alleviation, relaxation, and mood alternation (Tao et al., 2011; Wu et al., 2012). Therefore, researchers have argued that Chinese people tend to engage in gambling and report more gambling problems if they are less satisfied with life (Lai, 2006; Tang & Oei, 2011; Wong, Chan, Tai, & Tao, 2008). In the present study, we hypothesized that those who perceived a lower level of life satisfaction would report more symptoms listed in the DSM criteria for gambling disorder.

Demographics We also examined whether gender, age, educational attainment, and casino gambling are risk factors of gambling disorder in the present study. Previous research consistently demonstrated that Chinese men tend to report more symptoms of gambling disorder than Chinese women (Fong & Ozorio, 2005; Tao et al., 2011; Wong & So, 2003; Wu et al., 2012), but the impacts of age and education are far less conclusive. Age was found to be positively correlated with gambling frequency (Wu et al., 2013), but not symptoms of gambling disorder among Chinese people (Tao et al., 2011; Wong & So, 2003; Wu et al., 2012). For education, some studies showed that less education was associated with more 1 In the present study, these measurement items come from the existing criteria listed in the Chinese version of DSM–IV for pathological gambling (APA, 1994) because no Chinese version of DSM-5 (APA, 2013) is yet available.

WU, LAI, AND TONG

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

1192

symptoms of gambling disorder (Wong & So, 2003; Wu et al., 2013), but others found no such association (Fong & Ozorio, 2005; Tao et al., 2011; Wu et al., 2012). Game type preference may reflect gambling motives and expectations. For instance, social gambling, mainly driven by the socialization motive, does not contribute much to gambling disorder development among both Korean and Chinese gamblers (Fong & Ozorio, 2005; Lee, Chae, Lee, & Kim, 2007; Wu et al., 2012). In contrast, casino gamblers were consistently found to endorse more DSM criteria than noncasino gamblers in Macao (Fong & Ozorio, 2005; Tao et al., 2011; Wu et al., 2012). In the present study, we hypothesized that male gender, lower educational attainment, older age, and casino gambling would positively correlate with the number of DSM criteria endorsed.

The Aims of the Present Study The aim of the present study is therefore threefold: (a) to examine any differences in estimated prevalence of gambling disorder in Macao after application of the major changes on the DSM criteria and threshold, in order to evaluate the applicability of the changes in Chinese populations and provide the most up-todate prevalence rate of gambling disorder in Macao; (b) to evaluate the psychometric properties of the Chinese version of the nine DSM criteria as a common assessment tool of gambling disorder; and (c) to investigate the association between gambling disorder and some personal risk factors for a better understanding of this kind of mental disorder in a Chinese community with legalized gambling. To our best knowledge, the present study is the first empirical attempt to examine the prevalence rate of gambling disorder in a Chinese community with the consideration of the two major changes in the diagnostic criteria and threshold suggested by DSM-5. As criteria endorsement and interpretation may vary across groups such as culture (Alegría et al., 2009), the present study provides not only an updated prevalence rate of gambling disorder, but also a preliminary test on whether those changes on criteria and threshold are applicable to Chinese people; It thus facilitates further refinement of the DSM criteria as a diagnostic and screening tool in Chinese societies. Furthermore, this study aims to identify some risk factors of gambling disorder among Chinese gamblers in order to facilitate planning of effective prevention campaigns.

Method Respondents and Procedures After obtaining ethical approval from the affiliated university of the corresponding author, the telephone poll was conducted during January 26 –29, 2013. Using the computer assisted telephone interviewing procedure, we randomly sampled household telephone numbers from the 2012 Macao household phonebook and successfully interviewed 1,018 respondents. A person was considered eligible to this study if he or she was aged above 18, was a resident of Macau, and was able to speak Chinese. If multiple persons in a household were eligible, the one with the nearest birthday was selected. The participation of the respondents was totally voluntary and no monetary reward was involved. According to the American

Association for Public Opinion Research (2011), the cooperation rate (i.e., the proportion of all cases interviewed of all eligible participants ever contacted) was 59.7%. The interviews were conducted in either Cantonese or Mandarin Chinese by trained student assistants and each interview lasted about 10 minutes. The study was conducted in compliance with the American Psychological Association ethical principles. The sample consisted of 454 males (44.6%) and 564 females (55.4%) with a mean age of 41.6 years (SD ⫽ 17.1 years; range ⫽ 18 to 92 years). The majority of them were married (n ⫽ 621, 62.0%; as opposed to 36.3% single) and with a full-time job (n ⫽ 531, 52.8%). About one fifth of the respondents worked in casinorelated industries (n ⫽ 129, 21.2%). There was an oversampling of respondents with higher educational attainment, with 46 (4.6%) having kindergarten or below, 153 (15.4%) primary education, 197 (19.8%) junior secondary education, 286 (28.7%) senior secondary education, and 314 (30.8%) tertiary education. The median monthly income of the respondents was 11,250 Macau Patacas (MOP$; Roughly MOP$8 ⫽ US$1). Except for education, the demographic characteristics of the present sample closely matched those reported in the 2011 population census of Macao (Macao Statistics & Census Service, 2012b), taking into account that our sample included only those aged 18 years or above.

Measures Gambling behaviors and intention. The respondents were asked to indicate whether they had ever gambled prior to being interviewed (0 ⫽ no, 1 ⫽ yes), and if yes, whether they (i.e., gamblers) had ever gambled within the 12 months before the interview (0 ⫽ no, 1 ⫽ yes). Those who had gambled in the previous 12 months were considered recent gamblers in this study. This group of people were further asked to indicate what types of games (game type; namely casino gambling [including slot machines], soccer/basketball betting, horse racing, greyhound racing, mahjong house, social gambling [with families and/or friends at home], online casino games, Pacapio, casino ship gambling, sports betting with foreign bookmakers, and others) they had participated in, how much money they spent on gambling per gambling session (gambling expenditure), how much their maximum wager was (maximum wager size), and how often they gambled (gambling frequency), each by a single item. Gambling frequency was coded on a 9-point scale from 1-once a year to 9-every day. Intention to gamble was measured by another single item “I expect that I will be involved in gambling within the future 6 months,” with anchors 1-strongly disagree to 7-strongly agree. Gambling disorder. The Chinese version of the 10 criteria listed in DSM–IV for pathological gambling (APA, 1994) was administered for the present study. It has been commonly used to assess current prevalence of gambling disorder in Chinese samples (e.g., Fong & Ozorio, 2005; Wong & So, 2003; Wu et al., 2012). Only respondents who indicated recent gambling experience (i.e., in the past 12 months) answered these 10 self-report yes/no items (0 ⫽ no, 1 ⫽ yes). Two examples of the symptoms described a person who “is preoccupied with gambling” and “has repeated unsuccessful efforts to control, cut back, or stop gambling.” Cronbach’s alpha was .80 for the current sample. Materialism. The 6-item short form (Richins, 2004) of the Richins and Dawson’s (1992) material values scale were used to

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

GAMBLING DISORDER IN MACAO

measure respondents’ values with respect to the possession of material objects. Participants rated each item (e.g., “I admire people who own expensive homes, cars, and clothes.”) with a 7-point Likert scale, ranging from 1-strongly disagree to 7-strongly agree. The short form had an acceptable reliability (␣ ⫽ .71), and a higher score indicated stronger materialistic orientation. Life satisfaction. The Chinese version of the 5-item Satisfaction With Life Scale (Diener, Emmons, Larsen, & Griffin, 1985) was used to measure life satisfaction. The psychometric properties of the Chinese version were shown satisfactory by Sachs (2003). Participants responded with a 7-point Likert format, with anchors 1-strongly disagree to 7-strongly agree. In the present study the scale showed acceptable reliability (␣ ⫽ .77), and a higher score indicated higher life satisfaction.

Data Analysis We first examined the proportion of gamblers in our sample and their gambling and demographic profiles. Because some of the variables were ordinally scaled (e.g., education) or skewed (including income, gambling expenditure and wager size), they were analyzed by nonparametric statistical procedures, namely the chisquared tests and the Wilcoxon–Mann–Whitney tests. To understand the consequence of adopting the DSM-5 criteria, we compared the estimated prevalence with that computed using the DSM–IV criteria. Wilcoxon–Mann–Whitney tests and independent sample t tests were conducted to see, as opposed to the old criteria, how well the new criteria discriminated between disordered gamblers (DGs) and nondisordered gamblers (NDGs) based on other indicators of gambling disorder. Regarding the psychometric properties of 9-item DSM measure, we examined its factor structure, dimensionality (i.e., whether the items measure only one single construct), internal consistency, and its relationship with those indicators. Univariate chi-squared tests were performed to identify demographic risk factors of gambling disorder. Finally, we examined the psychological risk factors first by Spearman’s rank correlation analyses and then by a path model where life satisfaction and materialism were set to predict symptoms of gambling disorder, controlling for gender, age, education, and casino gambling. For the path analysis, a structural equation modeling approach with Mplus (Muthén & Muthén, 1998 –2012) was used to account for missing data, measurement errors of life satisfaction and materialism, and the categorical nature of the DSM items (with estimator ⫽ WLSMV).

1193

involved in casino gambling (Mdn ⫽ MOP$1,000) had statistically significantly higher gambling expenditure than noncasino gamblers (Mdn ⫽ MOP$100), U ⫽ 2458.5, z ⫽ ⫺7.91, p ⬍ .001, while social gamblers (Mdn ⫽ MOP$150) had statistically significantly less gambling expenditure than nonsocial gamblers (Mdn ⫽ MOP$350), U ⫽ 5503.5, z ⫽ ⫺3.38, p ⫽ .001. Female (51.8%) and single (54.5%) respondents were more likely to engage in social gambling than male respondents (38.6%), ␹2(1) ⫽ 4.98, p ⫽ .026, OR ⫽ 1.71, and married respondents (39.1%), ␹2(1) ⫽ 6.46, p ⫽ .011, OR ⫽ 1.87. On the other hand, married respondents (40.8%) were more likely to engage in casino gambling (20.9%), ␹2(1) ⫽ 11.96, p ⫽ .001, OR ⫽ 2.61, and so did respondents with lower educational attainment, Mann–Whitney U ⫽ 6932.5, z ⫽ ⫺2.67, p ⫽ .008. Employment status and personal income did not make statistically significant associations with the two game types (p ⬎ .05).

Estimating Prevalence of Gambling Disorder: DSM–IV Versus DSM-5

Gambling Profiles of the Sample

The prevalence rates of gambling disorder were estimated by the original 10-item criteria with five as cut-off and then by the 9-item criteria with four as cut-off. It should be noted that, among the recent gamblers (n ⫽ 282), one of them did not answer any of the DSM items, and was thus excluded from all subsequent analyses. Of the 10 DSM items answered by recent gamblers, the rate of missing responses ranged from 0.7% to 6.7%. The rate of respondents failing to respond to at least one item was 15.1%. DSM–IV criteria and cut-off. Ninety-one out of 281 recent gamblers (32.4%) met at least one of the 10 DSM–IV criteria. Given that some respondents missed at least one of the 10 DSM items, the subsequent analysis involved only the respondents of which we were confident to make DG or NDG classification based on their responses to the DSM items. Specifically, DGs respondents were those who scored 5 or above on only the items they had answered, and NDGs were those who scored 4 or below even if they endorsed all the items they missed.2 This resulted in a sample size of 272, among which 15 respondents were classified as DSM–IV DGs (based on 10-item criteria with 5 as cut-off), who accounted for 1.5% (95% CI [0.9%, 2.4%]; 12 males and 3 females) of all respondents in this study. Consideration of changes in DSM-5. In DSM-5, the item about Illegal Acts was dropped, and the cut-off for DG changed from 5 to 4. Strong and Kahler (2007) found the level of endorsement for this item was low, which was also true for our sample. Only one respondent in this study endorsed it, which was the lowest endorsement rate among all items. In addition, for this sample these two changes would lead to six more respondents

Among the 1,018 respondents, 509 (50.0%) said that they had participated in gambling prior to the interview (lifetime prevalence), of which 282 (27.7%, 95% Confidence Interval [CI] [25.0%, 30.5%]) said that had gambled within 12 months prior to the interview (past-year prevalence). Social gambling was endorsed the most (n ⫽ 128, 45.4%), followed by casino gambling (n ⫽ 95, 33.7%), mahjong house (n ⫽ 50, 17.7%), and lottery (n ⫽ 35, 12.4%). The median (Mdn) gambling expenditure per gambling session is MOP$200 (M ⫽ MOP$1,866, SD ⫽ MOP$8,323). Wilcoxon–Mann–Whitney tests revealed that those who had been

2 To check the influence of missing data on the DSM items, we performed sensitivity analyses by redoing the analyses using different treatments of missing data: (a) treating all nonresponses as absence of the corresponding symptoms, and (b) using Mplus with full information maximum likelihood and robust standard errors (estimator ⫽ MLR), and including gender, age, education, casino gambling (yes/no), and the observed DSM item responses as auxiliary variables (Enders, 2010; Graham, 2003). A respondent who did not respond to any of the 10 DSM items was excluded, which makes the sample size for both (a) and (b) as 281. The estimated prevalence rate with approach (a) was 1.5% (95% CI [0.9%, 2.4%]), and that with approach (b) was 1.8% (95% Wald CI [1.0%, 2.7%]).

Results

1194

WU, LAI, AND TONG

(total n ⫽ 21; 17 males and 4 females) to be categorized as DGs, making the prevalence rate of DGs increase to 2.1% (95% CI [1.4%, 3.2%]).3 The prevalence rate for males would then be 3.8%, and that for females would be 0.7%. If the cut-off score was kept as 5 in the 9-item DSM criteria, it would give the same classification of DGs as that from the DSM–IV criteria, because the only respondent who endorsed the illegal act item had scores changed from 3 to 2.

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Correlating With Other Indicators: DSM–IV Versus DSM-5 As shown in Table 1, the DSM-5 DGs (based on 9-item criteria with 4 as cut-off) had higher mean gambling intention than the DSM–IV counterparts (5.74 vs. 5.46), higher median gambling frequency (5.00 vs. 4.50), and lower median wager size (MOP$2,250 vs. MOP$3,000). It should be noted that the differences between DGs and NDGs became larger if the two changes in DSM-5 were considered. In particular, a significant difference between DGs and NDGs was found only on gambling expenditure and wager size with the DSM–IV criteria, whereas for DSM-5 DGs had statistically higher gambling intention, gambling frequency, gambling expenditure, and wager size than their counterparts (ps ⫽ .026, .002, ⬍.001, and ⬍.001), providing support that the recommended changes make the criteria and threshold more sensitive to disordered gambling patterns.4

Psychometric Properties of the DSM Criteria We used Mplus with the WLSMV estimator to conduct two exploratory factor analyses (EFAs), one 1-factor and one 2-factor, to check the dimensionality of the nine DSM items. The 1-factor model showed adequate fit to the data, with ␹2(N ⫽ 281, df ⫽ 27) ⫽ 27.78, p ⫽ .421, root mean squared error of estimation (RMSEA) ⫽ 0.01, comparative fit index (CFI) ⫽ 1.00, and weighted root mean squared residual (WRMR) ⫽ 0.06. The three largest eigenvalues for the sample correlation matrix were 5.29, 1.13, 0.75, so a single factor captured most of the generalized variance of the matrix. A likelihood ratio test comparing the 1-factor and the 2-factor model had ⌬␹2(df ⫽ 8) ⫽ 8.91, p ⫽ .350, so no statistically significant improvement was found for the 2-factor model. These results suggested that the unidimensionality assumption of the 9-item DSM measure is tenable. The internal consistency of the nine DSM items was .81, which was slightly higher than the 10-item version (␣ ⫽ .80) in our sample. Regarding the convergent validity, the DSM total score had significant positive correlation with gambling frequency (rank correlation ⫽ .14, p ⫽ .022), gambling expenditure (rank correlation ⫽ .36, p ⬍ .001), and wager size (rank correlation ⫽ .38, p ⬍ .001), but only a weak correlation with gambling intention (rank correlation ⫽ .09, p ⫽ .169).

Examining the Potential Risk Factors of Gambling Disorder Demographic factors. DSM-5 DGs (n ⫽ 21) were found more likely to be male (12.3% of males were DGs as opposed to 3.0% of females, ␹2(1) ⫽ 8.31, p ⫽ .004, OR ⫽ 4.57), and less well educated (M ⫽ 3.00 for DGs, 3.82 for NDGs), Mann–

Whitney U ⫽ 1515.5, z ⫽ ⫺2.97, p ⫽ .003. As hypothesized, similar analyses also found that DGs were more likely to be casino gamblers (16.9% of casino gamblers were DGs as opposed to 3.3% of noncasino gamblers, OR ⫽ 5.98, ␹2 ⫽ 15.49, p ⬍ .001). However, those who worked in casinos were no more likely to be DGs than those working in noncasino settings (8.9% vs. 8.0%, OR ⫽ 1.13, ␹2 ⫽ 0.38, p ⫽ .845). Also, no statistically significant differences were found between DGs and NDGs in terms of marital status, employment status, and personal income (ps ⬎ .17). Psychological factors. As hypothesized, materialism was found positively associated with the number of DSM criteria endorsed (range ⫽ 0 –9), with Spearman’s rank correlation ␳ ⫽ .24, p ⬍ .001. On the other hand, life satisfaction was found negatively associated with the DSM-5 total score, ␳ ⫽ ⫺.13, p ⫽ .039. To better interpret the results, we computed the standardized differences (i.e., Cohen’s d) of materialism and life-satisfaction between DGs and NDGs. DGs scored significantly higher on materialism, d ⫽ 0.51, and lower on life satisfaction, d ⫽ ⫺0.56 than their counterparts. Using structural equation modeling, we fitted a path model in which life satisfaction and materialism predict DSM-based symptoms of gambling disorder (see Figure 1). The outcome variable was the set of factor scores obtained when we fitted a Rasch model (see Embretson & Reise, 2000) to the nine DSM-5 items. As previously discussed, gender, age, education, and casino gambling were treated as controlled variables, which explained 14.2% of the variance of symptoms of gambling disorder. The fit-indexes suggested that the model had an acceptable fit to the data, with model ␹2(df ⫽ 86, N ⫽ 281) ⫽ 153.08, p ⬍ .001; RMSEA ⫽ 0.05, CFI ⫽ .91, and standardized root mean squared residual (SRMR) ⫽ 0.06. Casino gamblers scored higher on symptoms of gambling disorder than noncasino gamblers by an average of 0.51 SD (i.e., Cohen’s d effect size is 0.51, p ⬍ .001). Males also showed more symptoms of gambling disorder, d ⫽ 0.34 (p ⫽ .003). Age had a negative effect on symptoms (p ⫽ .031), whereas education had a positive effect (p ⬍ .001). Specifically, the model predicted that for one year increase in age the score of symptoms would decrease by 0.009 SD, whereas with each step increase in education (e.g., from primary to junior secondary) the score of symptoms would decrease by 0.19. When life satisfaction and materialism were entered, the explained variance of gambling disorder was 20.3% (a 6.1% increase). Materialism had a statistically significant positive effect on symptoms, ␤ ⫽ 0.25, p ⬍ .001, but the effect of life satisfaction was not statistically significant (␤ ⫽ 0.01, p ⫽ .851).

Discussion In this study with a representative Chinese adult sample in Macao, we observed that the lifetime prevalence of gambling participation was 50.0%. The estimated past-year prevalence rate was 27.7%, which was lower than rates among Macao residents in 2003 (67.9%), 2007 (59.2%), 2010 (55.9%), and 2011 (33.0%) (Fong & Ng, 2010; Wu et al., 2013). The statistics seem to indicate 3 The estimated prevalence rate with approach (a) was 2.1% (95% CI [1.4%, 3.1%]), and that with approach (b) was 2.1% (95% Wald CI [1.3%, 3.0%]). 4 In the additional sensitivity analyses, both approaches (a) and (b) supported the same conclusion.

GAMBLING DISORDER IN MACAO

1195

Table 1 Comparison of the Probable Disordered Gamblers (DGs) and NDGs 10 items

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

DG (ⱖ5) n (%) Gambling intention M (SD) Gambling frequency Mdn (SD) Gambling expenditure Mdn Maximum wager size Mdn

9 items NDG (⬍5)

15 (5.5%) 257 (94.5%) t(246) ⫽ 1.26, p ⫽ .210 5.46 (1.39) 4.85 (1.72) U ⫽ 1,134, z ⫽ ⫺1.91, p ⫽ .057 4.50 (2.52) 3.00 (1.96) U ⫽ 385, z ⫽ ⫺4.12, p ⬍ .001 3,000 100 U ⫽ 540.5, z ⫽ ⫺3.87, p ⬍ .001 3,000 200

DG (ⱖ4)

NDG (⬍4)

21 (7.7%) 251 (92.3%) t(246) ⫽ 2.24, p ⫽ .026 5.74 (1.24) 4.83 (1.73) U ⫽ 1,225, z ⫽ ⫺3.14, p ⫽ .002 5.00 (2.50) 3.00 (1.95) U ⫽ 414, z ⫽ ⫺4.80, p ⬍ .001 3,000 100 U ⫽ 596.5, z ⫽ ⫺4.46, p ⬍ .001 2,250 200

Note. N ⫽ 272, as 10 other respondents who had missing responses on the DSM items and whose statuses as DGs are not certain are excluded.

a decreasing trend in gambling participation among Macao residents but they must be viewed with caution because the variation can also be a result of the variation on sample size, sample characteristics, and definition of gambling involvement across studies. Among those recent gamblers, nearly half of them reported that they engaged in social gambling. This finding is consistent with the general observation that gambling is a socially acceptable activity for entertainment and socialization in Chinese communities (Loo, Raylu, & Oei, 2008). It is also not a surprise to find that, in a casino city like Macao, the second most frequent game type is casino gambling with about one third of the recent gamblers engaged in gambling in a casino. Due to the high acceptance and accessibility of gambling in Macao, the prevalence rate of gambling disorder is of concern. In the latest revision of DSM, the criterion of Illegal Acts was removed and the cut-off score for the diagnosis was lowered from 5 to 4. In our Chinese sample, the Illegal Acts criterion was the least frequently endorsed (n ⫽ 1) among 10 DSM criteria. Although the lowered

cut-off also resulted in six more respondents classified as DGs and the prevalence rate of gambling disorder increased from 1.5% to 2.1%, the classification accuracy was believed to increase. The present findings showed that DGs were better discriminated from NDGs in terms of other indicators of gambling disorder such as gambling frequency and spending, when the two changes in DSM-5 were adopted. If using all 10 DSM criteria and the old cut-off, DGs significantly differed from NDGs on only gambling wager size and expenditure. However, after removing the Illegal Acts criterion and using a cut-off of four, DGs were found not only to report higher intention to gamble but also to gamble more frequently with greater expenditure and wager size than NDGs. These findings indicate that the two recommended changes make the diagnosis method more sensitive to disordered gambling patterns in the Chinese community population. Similar to previous findings reported by studies conducted in the West (e.g., Denis et al., 2012; Strong & Kahler, 2007), our findings indicate that the 9-item DSM criteria, as an assessment tool for

Figure 1. Path model predicting the derived score on the 9-item DSM-5 gambling disorder measure. The DSM score is derived using Rasch scaling. The path coefficients are those when both variables linked by the path are standardized. Coefficients with ⴱ are statistically significant at .05 level. Males had a higher DSM score than females by 0.38 SD, whereas casino gamblers had a higher DSM score than noncasino gamblers by 0.44 SD.

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

1196

WU, LAI, AND TONG

gambling disorder, are unidimensional. The internal consistency of the set of criteria was high. Its convergent validity was also demonstrated by the correlation of the number of endorsed criteria with gambling frequency, wager size, and expenditure, in the expected direction. However, the DSM total scores only had weak and nonsignificant correlations with gambling intention. A possible reason is that the relationship between the DSM score and gambling intention is not linear, which was supported by the statistically significant differences in intention between DGs and NDGs. In sum, the present findings provide further evidence that the Chinese version of the DSM measure remains a valid and reliable measurement tool for gambling disorder among Chinese adults. Compared to life satisfaction, materialism was a more salient psychological risk factor of gambling disorder among Chinese adults in Macao. It was found to make an independent contribution to explaining the number of DSM criteria endorsed, after controlling for important demographics. Given that the motive of monetary gain was found positively correlated with materialism (Fang & Mowen, 2009) and gambling disorder symptoms (Tao et al., 2011; Wu et al., 2012), it is plausible that materialism may continually promote people’s gambling motive to win money in order to satisfy their materialistic need, thus making them vulnerable to gambling disorder. However, empirical assessment is needed for this proposed mechanism of mediation in future studies, particularly on Chinese casino gamblers, who often report money as their major motive to gamble (Tao et al., 2011; Wu et al., 2012). By examining the demographic characteristics of those probable disordered gamblers, other personal risk factors of gambling disorder in Macao were also identified. The hypotheses regarding gender and education, but not age, were supported. It was found that there was a significantly higher proportion of men to be classified as at-risk gamblers than that of women. Moreover, being male and having lower educational attainment also significantly accounted for the variance in gambling disorder symptoms. Men were more likely to be in community and clinical gambler samples than women in both Western and Chinese literature (e.g., Fong & Ozorio, 2005; Ibáñez, Blanco, Moreryra, & Sáiz-Ruiz, 2003; Tang, Wu, & Tang, 2007). Mental health professionals should pay additional attention to the psychosocial characteristics commonly shown by male disordered gamblers such as comorbid substance dependence (Ibáñez et al., 2003) and loss of interest at work (Tang et al., 2007). In addition, a recent local study found that education was associated positively with perceived control over gambling refusal and negatively with gambling involvement (Wu et al., 2013). This finding suggests that the welleducated people may have more psychosocial resources to handle negative affect (such as boredom) and to resist peer pressure to gamble than their counterparts, and thus they are less vulnerable to gambling disorder in Chinese societies. Consistent with previous local findings (Fong & Ozorio, 2005; Tao et al., 2011; Wu et al., 2012), casino gamblers reported more symptoms listed as DSM criteria than noncasino gamblers. Moreover, they spent more money on gambling than noncasino gamblers. Because casino gambling was vaguely defined as any gambling activities hosted at a casino in the current study, we could not evaluate the susceptibility of the gamblers of a specific casino game like baccarat. However, support from the casino corporates is still essential for effective gambling disorder prevention. In addition to offering selfexclusion and third-party exclusion programs, they can provide more training programs on gambling disorder to their staff. Moreover, some

kiosks with interactive design were set up in six Macao casinos and the users generally agreed that the information provided by those kiosks increased their understanding of and interest on responsible gambling in a pilot test (Macao Gaming Inspection & Coordination Bureau, 2013). Therefore, the coverage of this campaign can be extended to more casinos and more content about gambling disorder and treatment options can be added. To conclude, the present study found the Illegal Act criterion the least frequently endorsed item, and we recommended lowering the threshold for diagnosing gambling disorder in the Chinese adult population. With the cut-off of four out of nine, the estimated prevalence rate of gambling disorder was about 2.1% in Macao. Based on the findings of the risk factors identified, we suggested further intervention and research. However, several limitations of the present study should also be noted. First, the current telephone survey design not only imposed constraints on the survey questionnaire length but also excluded those Macao residents who do not have a household telephone line or cannot speak Mandarin or Cantonese Chinese. In order to obtain a satisfactory response rate, only the validated inventories of two psychological risk factors can be included in this study. Further studies with household survey design, which allows longer duration of an interview and more extensive coverage of the population, may complement and validate our findings. Second, due to its cross-sectional and observational design, the present study failed to test the causality among the variables. In addition, given that DSM-5 has been only recently published, the Chinese version of its criteria set of gambling disorder is not yet available. The present study only examined the impacts of two major changes but not those minor changes in the criteria of DSM-5. For example, the word often is added in some criteria to emphasize that only if a symptom often occurs, a particular criterion should be considered being met. Therefore, further studies should employ standard procedures of translation with the help from Chinese mental health professionals and further empirically test the measurement validity and reliability of Chinese version of DSM-5 criteria in both community and clinical samples in China. Data from clinical samples recruited in counseling or correctional service facilities may allow better evaluation of the efficacy of the Illegal Acts criterion because only one respondent endorsed with this item in our community sample in which only 21 disordered gamblers were identified. Finally, because our sample represented the general population in Macau, it only included a small number of potential disordered gamblers, and it is difficult to determine whether or not they had a gambling disorder without thorough assessment of their cognitions and behaviors. In order to demonstrate the classification accuracy of the DSM-5 criteria for Chinese gamblers, future studies with data on a larger clinical sample are encouraged.

References Alegría, A. A., Petry, N. M., Hasin, D. S., Liu, S. M., Grant, B. F., & Blanco, C. (2009). Disordered gambling among racial and ethnic groups in the US: Results from the national epidemiologic survey on alcohol and related conditions. CNS Spectrums, 14, 132–142. American Association for Public Opinion Research. (2011). Standard definitions: Final dispositions of case codes and outcome rates for surveys (7th ed.). Deerfield, IL: Author. Retrieved from http://aapor.org/Content/ NavigationMenu/AboutAAPOR/StandardsampEthics/StandardDefinitio ns/StandardDefinitions2011.pdf American Psychiatric Association. (1980). Diagnostic and statistical manual of mental disorders (3rd ed.). Washington, DC: Author.

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

GAMBLING DISORDER IN MACAO American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author. Denis, C., Fatséas, M., & Auriacombe, M. (2012). Analyses related to the development of DSM-5 criteria for substance use related disorders: An assessment of pathological gambling criteria. Drug and Alcohol Dependence, 122, 22–27. doi:10.1016/j.drugalcdep.2011.09.006 Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction with life scale. Journal of Personality Assessment, 49, 71– 75. doi:10.1207/s15327752jpa4901_13 Embretson, S. E., & Reise, S. P. (2000). Item response theory for psychologists. Mahwah, NJ: Erlbaum. Enders, C. K. (2010). Applied missing data analysis. New York, NY: Guilford. Fang, X., & Mowen, J. C. (2009). Examining the trait and functional motive antecedents of four gambling activities: Slot machines, skilled card games, sports betting, and promotional games. Journal of Consumer Marketing, 26, 121–131. doi:10.1108/07363760910940483 Fong, K. C., & Ng, M. P. (2010). Report of Macao Citizen’s Participation in Gambling Activities Survey 2010. Macao, China: Social Welfare Bureau of Macao Government. Retrieved from http://www.ias.gov.mo/ stat/rs/dfccvf_rs2010.pdf Fong, K. C., & Ozorio, B. (2005). Gambling participation and prevalence estimates of pathological gambling in a Far-East gambling city: Macao. UNLV Gaming Research & Review Journal, 9, 15–28. Graham, J. W. (2003). Adding missing-data-relevant variables to FIMLbased structural equation models. Structural Equation Modeling, 10, 80 –100. doi:10.1207/S15328007SEM1001_4 Ibáñez, A., Blanco, C., Moreryra, P., & Sáiz-Ruiz, J. (2003). Gender differences in pathological gambling. Journal of Clinical Psychiatry, 64, 295–301. doi:10.4088/JCP.v64n0311 Jiménez-Murcia, S., Stinchfield, R., Álvarez-Moya, E., Jaurrieta, N., Bueno, B., Granero, R., . . . Vallejo, J. (2009). Reliability, validity, and classification accuracy of a Spanish translation of a measure of DSM–IV diagnostic criteria for pathological gambling. Journal of Gambling Studies, 25, 93–104. doi:10.1007/s10899-008-9104-x Lai, D. W. L. (2006). Gambling and the older Chinese in Canada. Journal of Gambling Studies, 22, 121–141. doi:10.1007/s10899-005-9006-0 Lee, H., Chae, P. K., Lee, H., & Kim, Y. (2007). The five-factor gambling motivation model. Psychiatry Research, 150, 21–32. doi:10.1016/j .psychres.2006.04.005 Loo, J. M. Y., Raylu, N., & Oei, T. P. S. (2008). Gambling among the Chinese: A comprehensive review. Clinical Psychology Review, 28, 1152–1166. doi:10.1016/j.cpr.2008.04.001 Macao Gaming Inspection and Coordination Bureau. (2013). Report of the first-phase pilot test of responsible gambling information kiosk. Retrieved from http://www.dicj.gov.mo/web/files/responsible/ Kiosk%20testing%20report_20130927/Kiosk%20phrase%201_English .pdf Macao Statistics and Census Service. (2012a). Macao information yearly report 2012. Macao, China: Author. Macao Statistics and Census Service. (2012b). Results of 2011 population census. Macao, China: Author. Retrieved from http://www.dsec.gov.mo/ getAttachment/7a3b17c2-22cc-4197-9bd5-ccc6eec388a2/E_CEN_PUB_ 2011_Y.aspx Muthén, L. K., & Muthén, B. O. (1998 –2012). Mplus user’s guide (7th ed.). Los Angeles, CA: Muthén & Muthén. Mythily, S., Edimansyah, A., Qiu, S. J., & Munidasa, W. (2011). Quality of life in pathological gamblers in a multiethnic Asian setting. Annals Academy of Medicine Singapore, 40, 264 –268. Netemeyer, R. G., Burton, S., Cole, L. K., Williamson, D. A., Zucker, N., Bertman, L., & Diefenbach, G. (1998). Characteristics and beliefs associ-

1197

ated with probable pathological gambling: A pilot study with implications for the National Gambling Impact and Policy Commission. Journal of Public Policy & Marketing, 17, 147–160. Petry, N. M., Blanco, C., Auriacombe, M., Borges, G., Bucholz, K., Crowley, T. J., . . . O’Brien, C. (2014). An overview of and rationale for changes proposed for pathological gambling in DSM-5. Journal of Gambling Studies, 30, 493–502. doi:10.1007/s10899-013-9370-0 Reilly, C., & Smith, N. (2013). The evolving definition of pathological gambling in the DSM-5 (National Center for Responsible Gambling white paper). Washington, DC: National Center for Responsible Gaming. Retrieved from http://www.ncrg.org/sites/default/files/uploads/docs/white_papers/ ncrg_wpdsm5_may2013.pdf Richins, M. L. (2004). The material values scale: Measurement properties and development of a short-form. Journal of Consumer Research, 31, 209–219. doi:10.1086/383436 Richins, M. L., & Dawson, S. (1992). A consumer values orientation for materialism and its measurement: Scale development and validation. Journal of Consumer Research, 19, 303–316. doi:10.1086/209304 Sachs, J. (2003). Validation of the Satisfaction with Life Scale in a sample of Hong Kong university students. Psychologia, 46, 225–234. doi: 10.2117/psysoc.2003.225 Stinchfield, R. (2003). Reliability, validity, and classification accuracy of a measure of DSM–IV diagnostic criteria for pathological gambling. The American Journal of Psychiatry, 160, 180–182. doi:10.1176/appi.ajp.160.1.180 Strong, D. R., & Kahler, C. W. (2007). Evaluation of the continuum of gambling problems using the DSM–IV. Addiction, 102, 713–721. doi: 10.1111/j.1360-0443.2007.01789.x Tang, C. S. K., & Oei, T. P. (2011). Gambling cognition and subjective well-being as mediators between perceived stress and problem gambling: A cross-cultural study on White and Chinese problem gamblers. Psychology of Addictive Behaviors, 25, 511–520. doi:10.1037/a0024013 Tang, C. S., Wu, A. M. S., & Tang, J. Y. C. (2007). Gender differences in characteristics of Chinese treatment-seeking problem gamblers. Journal of Gambling Studies, 23, 145–156. doi:10.1007/s10899-006-9054-0 Tao, V. Y. K., Wu, A. M. S., Cheung, S. F., & Tong, K. K. (2011). Development of an indigenous inventory GMAB (Gambling Motives, Attitudes and Behaviors) for Chinese gamblers: An exploratory study. Journal of Gambling Studies, 27, 99–113. doi:10.1007/s10899-010-9191-3 Wong, I. L. K., & So, E. M. T. (2003). Prevalence estimates of problem and pathological gambling in Hong Kong. The American Journal of Psychiatry, 160, 1353–1354. doi:10.1176/appi.ajp.160.7.1353 Wong, V. W. K., Chan, E. K. M., Tai, S. P. K., & Tao, V. Y. K. (2008). Problem gambling among university students in Macao. Journal of Psychology in Chinese Societies, 9, 47–66. World Luxury Association. (2011, November). China is the second-largest country in luxury market, Hangzhou is the highest luxury consumption place in China. World Luxury Association. Retrieved from http://www.wla.hk/a/xwzx/ 2011/1101/3248.html Wu, A. M. S., Lai, M. H. C., Tong, K., & Tao, V. Y. K. (2013). Chinese attitudes, norms, behavioral control and gambling involvement in Macao. Journal of Gambling Studies, 29, 749–763. doi:10.1007/s10899-012-9344-7 Wu, A. M. S., & Tang, C. S. (2012). Problem gambling of Chinese college students: Application of the theory of planned behavior. Journal of Gambling Studies, 28, 315–324. doi:10.1007/s10899-011-9250-4 Wu, A. M. S., Tao, V. Y. K., Tong, K., & Cheung, S. F. (2012). Psychometric evaluation of inventory of Gambling Motives, Attitudes and Behaviors (GMAB) among Chinese gamblers. International Gambling Studies, 12, 331– 347. doi:10.1080/14459795.2012.678273

Received August 26, 2013 Revision received May 9, 2014 Accepted July 3, 2014 䡲

Gambling disorder: estimated prevalence rates and risk factors in Macao.

An excessive, problematic gambling pattern has been regarded as a mental disorder in the Diagnostic and Statistical Manual for Mental Disorders (DSM) ...
171KB Sizes 2 Downloads 6 Views