Prevalence of pathological gambling in treatment-seeking addicted patients: An exploratory study with the South Oaks Gambling Screen

In the current paper, the prevalence of pathological gambling in 112 treatment-seeking patients with substance addiction (81 alcoholics and 31 cocaine dependents) was estimated. The DSM-IV-TR diagnostic criteria for pathological gambling and the Spanish version of the South Oaks Gambling Screen (SOGS) were used. The results showed that 22.3% of substance-addicted patients had a comorbid diagnosis of pathological gambling. Furthermore, an additional 11.6% of the sample had relevant symptoms for problem gambling. In sum, 33.9% of the sample reported clinically significant gambling-related symptoms. From a socio-demographic point of view, all substance-addicted patients with gambling-related symptoms were men. A comparison between substance-addicted patients with and without pathological gambling showed significant differences in alcohol severity (assessed by the EuropASI), psychopathological symptoms (assessed by the SCL-90-R) and personality variables (assessed by the MCMI-II). In all cases, scores were significantly higher in gamblers than in non-gamblers. Finally, the implications of these results for further research and clinical practice are commented upon.


Introduction
In recent years, there has been a growing interest in the study of comorbid disorders in addictive behaviours.This is probably due to the high rate of comorbidity observed in substance abusers and its large influence on therapeutic results.Efforts to improve the success rates of intervention programs for addicted patients have generated a clear interest in the study of variables that interfere with the results, or limit the reach, of these programs (Fernández-Montalvo et al., 2004).
The aims of this study were to carry out an accurate assessment of pathological gambling in a sample of addicted patients seeking outpatient treatment and to compare the severity of gambling dependence, as well as the psychopathological and personality variables, between addicts with and without pathological gambling.For the diagnosis of pathological gambling, DSM-IV-TR criteria (American Psychiatric Association, 2000) and the Spanish version (Echeburua, Baez, Fernández-Montalvo & Páez, 1994) of the South Oaks Gambling Screen (SOGS) (Lesieur & Blume, 1987) were used.This Spanish version has good psychometric properties.For the assessment of psychopathological and personality variables the SCL-90-R (Derogatis, 1992) and the MCMI-II (Millon, 1997) were used, because they are the more frequently used in other studies.

Method Participants
The sample consisted of 112 substance-addicted patients (81 alcoholics and 31 cocaine dependents) who sought outpatient treatment at the "Proyecto Hombre Addiction Treatment Program" in Pamplona, Spain, during the period from November 2006 to August 2008.The assessment was made as a standardized protocol to all patients seeking for the treatment program.According to the current study's admission criteria, the patients had to a) meet the diagnostic criteria of alcohol or cocaine dependence according to DSM-IV-TR (APA, 2000); b) be between 18 and 65 years old; and c) give their informed consent to participate in the study.All eligible patients agreed to participate in the study.

Addiction variables
The EuropASI (Kokkevi & Hartgers, 1995) is the European version of the Addiction Severity Index (ASI) (McLellan, Luborsky, O'Brein & Woody, 1980).This interview assesses the need for treatment in the following six areas: a) general medical state; b) labour and economic situation; c) drug consumption (alcohol included); d) legal problems; e) family and social relationships; and f) psychological state.Scores range from 0 (no problem) to 9 (extreme problem) in each area and the cut-off point for each area is 4. The Spanish version was used in this study (Bobes, Gónzález, Sáiz, & Bousoño, 1996).

Gambling variables
The SOGS (Lesieur & Blume, 1987; Spanish version by Echeburúa et al., 1994) is a screening questionnaire composed of 20 items that are related to gambling behaviour, loss of control, sources of money and emotions involved.In the Spanish version, this assessment tool has a test-retest reliability of .98 and an internal consistency of .94.The convergent validity with DSM-IV criteria is .92.The range is from 0 to 19.A score higher than 4 (the cut-off point) serves to identify probable pathological gamblers.

Psychopathological variables
The Symptom Checklist-90-Revised (SCL-90-R) (Derogatis, 1992; Spanish version of González de Rivera, 2002) is a selfadministered general psychopathological assessment questionnaire.It comprises 90 items with 5 alternatives for each on a Likert-type scale, ranging from 0 (none) to 4 (very much).The aim of the questionnaire is to indicate a subject's symptoms of psychological disturbance.As it has been shown to be sensitive to therapeutic change, it may be used for either single or repeated assessments.The SCL-90-R consists of nine areas of primary symptoms: somatisation, obsessive-compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation and psychoticism.It also provides three indices that reflect the subject's overall level of severity.The Millon Clinical Multiaxial Inventory (MCMI-II) (Millon, 1997; Spanish version by Ávila, 2002) is a 175-item, true/false, self-report questionnaire.It was designed to identify clinical states and personality disorders similar to those contained in the DSM.The MCMI-II contains eight basic personality scales: 1) Schizoid-asocial; 2) Avoidant; 3) Dependent-submissive; 4) Histrionic-gregarious; 5) Narcissistic; 6) Antisocialaggressive; 7) Compulsive-conforming; and 8) Passiveaggressive.In addition to the basic personality patterns there are three pathological personality scales: Schizotypal (S), Borderline (B) and Paranoid (P).The MCMI-II also contains nine symptom scales: Anxiety, Somatoform, Hypomanic, Dysthymic, Alcohol abuse, Drug abuse, Psychotic thinking, Psychotic depression, and Psychotic delusion.However, in this study these additional Axis I clinical syndrome scales have not been taken into account because they are not relevant to the purposes of this research.

Procedure
Once the clinical sample was selected according to the previously described criteria, assessment of the sample was carried out in three sessions.In the first one, data related to socio-demographic characteristics and to drug consumption were collected.In the second session, the presence of psychopathological symptoms (with the SCL-90-R) and pathological gambling (with the SOGS) were assessed.Finally, in the third one, prevalence of personality disorders was assessed using the MCMI-II.According to the conservative criteria of Weltzler (1990) regarding the MCMI-II, a base rate score above 84 is considered to be significant.

Prevalence of pathological gambling
Results showed that 22.3% (25 subjects) of the total sample met the diagnostic criteria of DSM-IV-TR (APA, 2000) for pathological gambling.This rate of comorbidity corresponded with results obtained with the SOGS.
The cut-off point in the Spanish version of the SOGS is four points.The mean SOGS score for the entire addiction sample was 2.06 (SD = 3.3).However, from a clinical perspective, the number of addicted patients with a score above 4 was 25 subjects (22.3% of the total sample).An additional 11.6% (n = 13) endorsed two or three symptoms, which, in the Spanish version is considered an indication of a potential or preclinical gambling problem.These figures show that 33.9% of this sample reported clinically significant gambling-related symptoms (22.3% with pathological gambling and 11.6% with problem gambling).
The rate of pathological gambling in alcoholic patients was 23.4% (19 patients of the alcoholic sample), higher than that observed in cocaine addicted patients (6 cases; 19.3%).However, there was no significant difference between the two samples (X 2 = 0.22; p = .64).

Comparison between groups
From a socio-demographic point of view, all addicted patients with pathological gambling were men.No addicted women showed an associated gambling problem.
Results of a comparison between patients with and without pathological gambling in addiction severity, as measured by the EuropASI, are presented in Table 1.As can be seen, the only statistical difference between groups was observed in the Alcohol severity area, which was significantly higher in addicts with gambling than in addicts without gambling.In the rest of variables of the EuropASI there were no statistical differences between gamblers and non-gamblers.SCL-90-R results regarding psychopathology are shown in Table 2. Statistical differences between groups were observed in most of the 12 scales of the SCL-90-R.Except for Hostility scale, comparisons between both groups showed that scores of the gamblers were situated significantly above scores of the non-gambler addicts.
anales de psicología, 2012, vol.28, nº 2 (mayo) Finally, from a quantitative point of view, statistical differences were observed in most of the scales of the MCMI-II (Table 3).With the exception of the compulsive scale the scores of gambler addicts were significantly higher than of nongamblers.However, from a qualitative point of view, there was only one statistical difference between groups in the rate of personality disorders (Table 4).Addicts with pathological gambling had a more prevalent rate of passive-aggressive personality disorder.

Discussion
This research contributes further evidence of an elevated prevalence rate of pathological gambling among addicted patients (Petry, 2007).As in previous studies (Cunningham-Williams et al., 2000;Daghestani et al., 1996;Fernández-Montalvo et al., 2005;Langenbucher et al., 2001;Lesieur et al., 1986;Orford et al., 2003;Rodríguez-Martos, 1989;Sellman et al., 2002;Welte et al., 2004), pathological gambling has a higher presence among addicted patients than in the general population (about 2% in epidemiologic studies).In summary, results of this study show that 22.3% of the sample presented an associated problem of pathological gambling (23.4% of alcoholics and 19.3% of cocaine addicts).All of them are men.This coincides with the results found in previous studies, which show that few women with a pathological gambling problem seek treatment in clinical settings (Echeburúa, González, Corral y Polo, 2011).Moreover, an additional 11.6% obtained a SOGS (Lesieur & Blume, 1987) score that reflected the presence of problem gambling.These data show that 33.9% of this sample reported clinically significant gambling-related symptoms.
This figure is worrying, because most of the standard programs for clinical intervention with addicted patients do not include an assessment of pathological gambling.However, the results of this study show a need to take into account problems related to gambling, both in the clinical evaluation of addictions and in the development of specific treatment programs.Otherwise, the gambling problem could remain hidden or overlapped with the drug dependence and hinder therapeutic intervention.
Moreover, the risk of developing a substitute dependence in the treatment of the addictive behaviours should be considered.In these cases, the treatment of drug addiction may exacerbate the problems of pathological gambling presented by the patient or precipitate a relapse in drug consumption once abstinence is achieved.Therefore, some authors have suggested the possibility of a combined treatment for drug dependence and pathological gambling for this type of patient, with the aim of preventing the development of an eventual substitute dependence (Daghestani et al. 1996;Fernández-Montalvo et al., 2005;Rodríguez-Martos, 1989).
On the other hand, the comparison between addicts with and without an associated problem of pathological gambling showed the existence of significant differences in psychopathological and personality variables.The results show an increased presence of psychopathological comorbidity in addicted patients who have an associated problem of gambling.These results agree with those obtained in previous studies (cf. Fernández-Montalvo et al. 2005;Langenbucher et al., 2001;Petry, 2000;Steinberg et al., 1992;Villas et al., 2005).All of them reflect a greater severity in addicted patients with pathological gambling.
Nevertheless, a limitation of this study is that the presented results were obtained with a relatively small sample (112 addicts).Therefore, it would be more appropriate to replicate this research with a wider sample of addicted patients, and with a bigger variety of drug abuse, not only alcohol and cocaine.Also, data from this study do not indicate the temporal relationship between the two addictions.In this sense, a logistic regression analysis could be suitable to test this kind of relationship, as well as to achieve a complete framework of the relationships between pathological gambling and other kind of variables: age, gender, kind of addiction (alcohol vs. cocaine), SCL-90-R and MCMI-II scores.In the study by Rodríguez-Martos (1989), 57% of patients believed that the problem of alcoholism preceded the problem of pathological gambling.Also, Cho et al. (2002) found that alcohol problems preceded gambling in most dually diagnosed alcohol-dependent men.However, Cunningham-Williams et al. (2000) found that most pathological gamblers began using cigarettes, marijuana and alcohol prior to developing gambling problems, but pathological gambling often preceded dependence on other drugs, especially stimulants.This is an interesting question for future research.

Table 3 .
Comparison in the MCMI-II.

Table 4 .
Personality disorders with the MCMI-II.The total number of people affected by personality disorders is less than the total sum of disorders because there are patients who presented more than one personality disorder.