Sub-typing Gamblers on the Basis of Affective Motivations for Gambling: Implications for Treatment Matching Dr. Sherry H. Stewart, Departments of Psychiatry

Sub-typing Gamblers on the Basis of Affective Motivations for Gambling: Implications for Treatment Matching Dr. Sherry H. Stewart,

Departments of Psychiatry & Psychology,

Dalhousie University,

CIHR Investigator,

Killam Research Professor

Funding

Ontario Problem Gambling Research Centre (OPGRC)

Nova Scotia Gaming Foundation (NSGF)

Collaborators

Dr. Sean Barrett Dr. Joel Katz

Dr. Raymond Klein Dr. David Hodgins Dr. Anne-Marie Wall Dr. Martin Zack

Pamela (Loba) Collins Fofo Fragopoulos Adrienne Girling Meghan Kirsch Melissa Mohan Sarah Stuart

Background: Motivational Theories Motivational models of addiction argue that people

engage in addictive behaviors to obtain desired outcomes (e.g., Cooper, 1994)

Many such theories point to desires for mood alteration as motivating addictive behavior (e.g., Cox & Klinger, 1988)

Theories abound regarding contributions of motivations involving emotional self-regulation in etiology and maintenance of pathological gambling Desire for tension-/dysphoria-reducing effects (e.g.,

Beaudoin & Cox, 1999) Desire for euphoric consequences (e.g., Hickey et al.,

1986)

Background: DSM-IV-TR Background: DSM-IV-TR Criteria for Pathological Criteria for Pathological GamblingGambling

A. Persistent and recurrent maladaptive gambling behavior A. Persistent and recurrent maladaptive gambling behavior as indicated by five (or more) of the following:as indicated by five (or more) of the following:

(1) is preoccupied with gambling(1) is preoccupied with gambling

*(2) needs to gamble with increasing amounts of money in *(2) needs to gamble with increasing amounts of money in order to achieve the order to achieve the desired excitementdesired excitement

(3) has repeated efforts to control, cut back, or stop (3) has repeated efforts to control, cut back, or stop gamblinggambling

(4) is restless or irritable when attempting to cut (4) is restless or irritable when attempting to cut down or stop gamblingdown or stop gambling

Background: DSM-IV-TR Background: DSM-IV-TR Criteria for Pathological Criteria for Pathological GamblingGambling

**(5) gambles as a way of escaping from problems or of (5) gambles as a way of escaping from problems or of relieving a dysphoric moodrelieving a dysphoric mood

(6) after losing money gambling, often returns another day to (6) after losing money gambling, often returns another day to get even ("chasing" one's losses)get even ("chasing" one's losses)

(7) lies to family members, therapist, or others to conceal the (7) lies to family members, therapist, or others to conceal the extent of involvement with gamblingextent of involvement with gambling

(8) has committed illegal acts such as forgery, fraud, theft, or (8) has committed illegal acts such as forgery, fraud, theft, or embezzlement to finance gamblingembezzlement to finance gambling

(9) has jeopardized or lost a significant relationship, job, or (9) has jeopardized or lost a significant relationship, job, or educational or career opportunity because of gamblingeducational or career opportunity because of gambling

(10) relies on others to provide money to relieve a (10) relies on others to provide money to relieve a

desperate financial situation caused by gamblingdesperate financial situation caused by gambling

Background: Heterogeneity

Increasing recognition of the heterogeneity of gamblers (e.g., Ferris, & Wynne, 2001)

Cluster analytic studies suggest subtypes For example, Blaszczynski & Nower (2002)

suggest three clusters: Antisocial impulsivist gamblers Emotionally vulnerable gamblers Behaviorally-conditioned gamblers

Researchers Proposed SubtypesLesieur Impulsive Impulsive(2001) (2 of 3) Escape Seekers Action SeekersMcCormick Recurrently Chronically (1987) Depressed Under-StimulatedBlaszczynski et al. Depression- Boredom-(1990) Prone ProneBlaszczynski & Nower Emotionally Antisocial(2002) (2 of 3) Vulnerable Impulsivist Grant Anxious/ Pleasure/Urge &(2007 – yesterday’s Depressed/ General Impulsivity talk) Obsessional Need for Stimulation

Previous Sub-Typing Schemes

Background: Sub-typing Substance Abusers Our previous work in the alcohol/drug

abuse area involves sub-typing substance abusers according to underlying motives for substance use (Conrod et al., 2000a)

This sub-typing predicts: Substance use patterns/preferences Co-morbid psychopathology Response to “matched” interventions

(Conrod et al., 2000a,b, 2006; Watt et al., 2006)

Matching brief interventions to motivational profiles

Random assignment to 1 of 3 90-minute interventions:

(1) Motivation-matched cognitive-behavioral training (N=94)

(2) Motivation-mismatched cognitive-behavioral training (N=97)

(3) Film control (N=52)

(Conrod, Stewart et al., 2000; Psych of Addictive Behaviors)

Procedure: Follow-up

Assessment at 6-months post-treatment Telephone interview Interviewer blind to subtype and intervention Several substance-related outcomes assessed

0

10

20

30

40

50

60

Matched Mismatched Film

% r

em

itte

dSix-months Remission Rates

0

1

2

3

4

5


Re

du

cti

on

in

de

pe

nd

en

ce

sym

pto

ms

Reduction in Dependence Symptoms

0

5

10

15

20

25


% r

ep

ort

ing

>6

0 d

ays

ab

sti

ne

nt

fro

m a

lco

ho

lLengthy Abstinence from Alcohol

Our New Proposal:

What about sub-typing gamblers according to motivations for emotional regulation?

Might be a theoretically- and clinically-useful way of understanding diversity among pathological gamblers and those at risk.

Study 1: Sub-Typing Problem Gamblers Who Drink When Gambling on Affective Motives for Gambling

Background: Gambling and Alcohol High co-morbidity between pathological

gambling and alcohol use disorders (Ciarrocchi, 1993; Crockford & el-Guebaly, 1998; Griffiths, 1994; Stewart & Kushner, 2003).

Theories explaining co-morbidity include: Pathological Gambling Alcohol Abuse (Zack et al., 2005) Alcohol Abuse Pathological Gambling (Ellery et al., 2005) Common Third Variable

For example, common underlying motivations for gambling/drinking

Purposes

Test the validity of a gambler subtyping scheme that classifies gamblers on their primary emotion regulation motives for gambling

Test the degree of overlap in specific motives for gambling and for drinking among problem gamblers who drink when gambling Do those who use gambling to cope with negative emotions

also use drinking for the same reason? Do those who use gambling to enhance positive states also

use alcohol for the same reason?

Hypotheses

(1) At least two clusters or subtypes of pathological gamblers would emerge using gambling situations as indirect measure of motives those gambling in response to “negative” contexts (e.g.,

unpleasant emotional states) – COP gamblers. those gambling in response to “positive” contexts (e.g., pleasant

emotional states) – ENH gamblers. (2) Clusters of gamblers identified using the Inventory of

Gambling Situations (indirect) would be validated using Gambling Motives Questionnaire (direct).

(3) The ENH gamblers would evidence greater gambling behavior than other cluster(s)

(4) The COP gamblers would evidence greater gambling problems than other cluster(s)

Hypotheses (continued)

(5) We expected to observe associations between gambling motives and drinking motivations on the Drinking Motives Questionnaire

(6) We hypothesized that the ENH gamblers would show higher quantities of drinking, and more alcohol related problems on the Brief Michigan Alcoholism Screening Test, than other cluster(s)

Method - Participants

A sample of 158 gamblers (77%M; mean (SD) age = 36.0 (10.7) years) was recruited from the community using newspaper and television advertisements.

All were probable pathological gamblers (scoring > 5 on the SOGS).

All reported consuming alcohol at least 50% of the times when they gamble.

All were 19 years of age or older

Results: Hypothesis 1aPattern Matrix from Principal Components Analysis of subscale scores from the Inventory of Gambling

Situations (IGS; Turner & Littman-Sharp, 2006) following Oblique Rotation (n = 158)

Factor 1 – Factor 2 – IGS Subscale Negative Situations Positive Situations CommunalityUnpleasant Emotions .976* -.029 .918Worried Over Debt .945* -.072 .812Conflict With Others .938* -.034 .840

Testing Personal Control .886* .032 .821Winning and Chasing Losses .764* .202 .821

Pleasant Emotions -.105 .930* .751 Social Pressure -.036 .899* .768Need for Excitement .246 .757* .872Confidence in Skills .173 .678* .639

Urges and Temptations .603* .423* .873

Note: A two-factor solution was selected based on Kaiser’s rule. Salient loadings (> .40) are indicated with an asterisk (*).

Results: H1b: Cluster Analysis 3 cluster solution Cluster 1 – high positive situation scores/low

negative situation scores – labeled Enhancement (ENH) gamblers (n = 94)

Cluster 2 – high negative/high positive situation scores, but particularly high negative scores – labeled Coping (COP) gamblers (n = 36)

Cluster 3 – low negative/low positive situation scores – labeled Low Emotion Regulation (Low ER) gamblers (n = 28)

Results: H2: Factor Scores on Gambling Motives Questionnaire (GMQ) by Subtype

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

ENH COP Low ER

coping motivesenhancement motives

cluster group x subscale interaction (F (2, 151) = 3.44, p < .05)

Results: H3: Number of Lifetime Gambling Activities (SOGS) by Subtype

0

1

2

3

4

5

6

7

8

9

10

ENH COP Low ER

# activities

cluster group effect (F (2, 155) = 19.90, p < .001)

a a b

Results: H4: Gambling Problems on DSM-IV-Based Measure by Subtype

0

2

4

6

8

10

12

14

16

18

20

ENH COP Low ER

gambling problems


ab c

Results: H5: Factor Scores on Cooper’s (1994) Drinking Motives Questionnaire (DMQ) by Subtype

-1

-0.5

0

0.5

1

1.5

ENH COP Low ER

coping motivesenhancement motives

cluster group x subscale interaction (F (2, 151) = 7.42, p < .005)

Results: H6a: Drinking Quantity (drinks per occasion) by Subtype

0

1

2

3

4

5

6

7

8

9

10

ENH COP Low ER

drinking quantity


aa b

Results: H6b: Drinking Problems by Subtype

0

2

4

6

8

10

12

14

ENH COP Low ER

B-MAST scores


ab b

Some Remaining Questions

Do the clusters represent trait-like subtypes or stages?

Do the clusters simply reflect correlates of gambling problem severity?

Are the sub-typing findings specific to pathological gamblers who drink when gambling?

Study 2: The Relationship of Gambler Subtypes to Gambling Expectancies in Regular Gamblers

Background: Expectancies

Expectancies = beliefs about consequences of engaging in a specific behavior

“If-then” statements making connections between behavior and expected outcome (Goldman et al., 1999)

Distinct from motives (Birch et al., 2004) Many domains, but two higher-order

Reward Relief

(Birch et al., 2004)

Background: Expectancies

In addictions field, expectancies exert powerful influence on addictive behavior e.g., expectancies about effects of alcohol strong

determinants of heavy and problem drinking (Goldman et al., 1999)

Surprisingly, little research on expectancies in gambling area

Purposes

Replicate typology of gamblers identified in Study 1, with sub-typing on the basis of gambling situations (inferred motives)

Determine if gambler subtypes differ in gambling outcome expectancies on a new mood-regulation Gambling Expectancies Questionnaire (GEQ) modeled after a similar alcohol outcome expectancies measure (Birch et al., 2004; Singleton et al., 1994)

Relief expectancies (e.g., “I would feel less irritable if I gambled now”)

Reward expectancies (e.g., “It would be great to gamble now”)

Hypotheses

(1) Three clusters or subtypes of regular gamblers would emerge on basis of IGS positive and negative situation factors: Pure enhancement gamblers (ENH) Primarily coping gamblers (COP) Low emotion-regulation gamblers (LOW ER)

(2) Gambler subtypes would differ in mood-regulation gambling outcome expectancies on the GEQ Reward expectancies: ENH = COP > LOW ER Relief expectancies: COP > ENH = LOW ER


A sample of 181 regular gamblers (56% M; mean age = 37.7 years) was recruited from the community using newspaper and television advertisements.

Approximately half recruited in Halifax (n=84) and the other half in Toronto (n=97)

No restrictions regarding SOGS scores or alcohol consumption; M SOGS = 3.88 (4.23)

Results: Hypothesis 1aPattern Matrix from Principal Components Analysis of subscale scores from the IGS following Oblique Rotation (n = 179)

Factor 1 – Factor 2 – IGS Subscale Negative Situations Positive Situations CommunalityConflict With Others .938* -.093 .789Worried Over Debts .904* -.100 .724Testing Personal Control .794* .130 .766Unpleasant Emotions .786* .144 .768

Need To Be In Control .775* .141 .745

Pleasant Emotions -.201 .952* .727 Social Pressure .029 .750* .588Need for Excitement .192 .725* .720Confidence in Skills .205 .704* .702

Urges and Temptations .403* .580* .766Winning and Chasing Losses .448* .500* .707


Results: Hypothesis 1bGambling situation factor scores of the three clusters of regular gamblers

-1.5

-1

-0.5

0

0.5

1

1.5

Cluster 1 ENH(n=68)

Cluster 2 LOWER (n=63)

Cluster 3 COP(n=48)

Negative SituationsPositive Situations

Results: Hypothesis 2GEQ factor scores as a function of gambler subtype

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

Cluster 1 ENH(n=67)

Cluster 2 LOWER (n=61)

Cluster 3 COP(n=46)

Relief ExpectanciesReward Expectancies

cluster group x expectancy domain interaction (F (2, 171) = 25.80, p < .001)

STUDY 2 Cluster 1 ENH

Cluster 2 LOW ER

Cluster 3 COP

Age (yrs) 33.57 (14.54)b

41.02 (14.02)a

39.88 (9.03)a

SOGS 2.99 (2.45)b

1.41 (2.12)c

8.48 (4.79)a

Gambling freq (per wk)

2.40 (1.96)a

1.57 (1.02)b

2.84 (1.96)a

B-MAST 4.28 (5.70)b

2.40 (3.97)c

7.58 (6.34)a

Drinking Onset (yrs)

14.91 (2.90)

15.42 (2.67)b

14.00 (2.62)a

Study 3: Validity of Sub-Typing Undergraduate Gamblers on Affective Motives for Gambling

Purposes

Test the validity of a gambler subtyping scheme that classifies gamblers on their primary emotion regulation motives for gambling, as applied to undergraduate gambling behavior

Does the three subtype scheme hold in undergraduate gamblers?

Do the subtypes vary as expected on gambling behaviors and gambling problems?


A sample of 168 regular gamblers (58% M; mean age = 20.3 years) was recruited from the undergraduate student population

Approximately half recruited at Dalhousie University (n = 82) and the other half recruited at York University (n = 86)

No restrictions regarding gambling problems or alcohol consumption

Results: Hypothesis 1aPattern Matrix from Principal Components Analysis of subscale scores from the IGS following Oblique Rotation (n = 168 university student gamblers)

Factor 1 – Factor 2 – IGS Subscale Positive Situations Negative Situations

Pleasant Emotions 1.101* -.234Social Pressure .818* .025Confidence in Skills .796* .025

Need for Excitement .790* .146

Conflict with Others -.071 .982*Negative Emotions .044 .900*Worried about Debt -.027 .886*Testing Personal Control .241 .761*

Urges and Temptations .597* .456*Winning and Chasing Losses .580* .401*


Results: H1b: Gambling Situation Factor Scores of the 3 Clusters of Gamblers

1=cop/enh, 2=enh, 3=low er321

Mea

n

4

3

2

1

0

-1

-2

Gambling Situation Factors Scores for the Three Clusters of Regular Undergraduate Gamblers

negative situationspositive situations

1 = COP (n=22); 2 = ENH (n=65); 3 = Low ER (n=81)

STUDY 3 Cluster 1 COP

Cluster 2 ENH

Cluster 3 Low ER

Age (yrs) 22.32 (4.93)a

19.88 (1.96)b

20.08 (2.38)b

Gender (% male)*

81.8% 61.5% 48.2%

Money Spent When Gambling ($)

50.20 (70.04)a

37.22 (46.86)a

17.82 (21.46)b

Hours per week Gambling

6.81 (6.95)a

3.06 (2.74)b

2.00 (1.78)c

Results: Percentage High Risk Gamblers (on CPGI) as a function of Subtype

1 = COP (n=22); 2 = ENH (n=65); 3 = Low ER (n=81)CPGI group

highlow

Per

cen

tag

e o

f ea

ch c

lust

er

100.0%

80.0%

60.0%

40.0%

20.0%

0.0%

321

1=cop/enh, 2=enh, 3=low er

Percent of Each Cluster Scoring Low and High on the CPGI

Cluster

Study 4: Heart Rate Increase to Alcohol and VLT Play as a Function of Gambler Subtype

Background: Heart Rate & Reward Alcohol intake leads to resting heart rate

increases (e.g., Conrod et al., 2001; Stewart et al., 1992)

Positively related to feeling energized in regular drinkers (Conrod et al., 2001)

Suggested marker for susceptibility to alcohol reward and reinforcement (e.g., Brunelle et al., 2003; Peterson et al., 1993)

Background: Heart Rate & Reward Heart rate increases may apply to a variety of

addictive behaviours Heart rate increases have also been seen in

gamblers when gambling (e.g., Coventry & Norman, 1997; Meyer et al., 2004)

Two prior studies show heart rateincreases to alcohol and gambling alone and in combination in regular gamblers (Stewart et al., 2005, 2006).

Experimental Design

2 x 3 x 3 mixed model

Between subjects factors: Beverage condition: alcohol vs. placebo control Gambler subtype: COP vs. ENH vs. Low ER

Within subjects factor Testing time: pre-drink vs. post-drink vs. VLT play

(spinning reels)

Participants

72 regular VLT players Recruited from the community Mean age: 35 years old 38 males; 34 females 47 NPG; 25 PPG (SOGS) Play VLTs 1.9 times per week,

on average

IGS Cluster AnalysisIGS Cluster Analysis

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

3

Negative Situations

Positive Situations

ENH(N=36)

LOW ER(N=31)

COP(N=5)

Characteristics of Subtypes

SOGS total scores 4.97 (3.84)* 2.03 (1.85)*

Gambling frequency (per week)Drinking frequency (per week)Number of drinks per week 11.98 (1.24)* 7.87 (1.34)*Number of drinks while gambling 4.60 (3.64)* 2.74 (2.61)*

2.36 (1.51)* 1.56 (.79)*

Enhancement (ENH)

Low Emotion Regulation (LOW ER)

2.26 (1.27)* 1.45 (.92)*

p < 0.02

Heart Rate Analyses

DV = % change from pre-drink (sober) baseline

ANOVA revealed hypothesized three-way interaction (p = .036)

Beverage Condition X Subtype X Testing Time

Follow-up simple effects analyses for heart rate response at Post-Drink and VLT play

VLT Play

p = .03

0%

2%

4%

6%

8%

10%

12%

14%

16%

ENH LOW ER

He

art

Ra

te P

erc

en

t C

ha

ng

e f

rom

Pre

-d

rin

kin

g B

as

eli

ne

Placebo

Alcohol

p = .03

Studies 1-3: Conclusions Three subtypes of gamblers were consistently identified based on

cluster analyses of IGS scores and validated with the GMQ: a group that gambles purely for positive reinforcement (pure enhancement;

ENH) a group that gambles for both positive and negative reinforcement, but

mainly for negative reinforcement (coping; COP) a group that gambles for neither positive nor negative reinforcement (low

emotion-regulation; LOW ER) This subtyping scheme applies to

Pathological gamblers who drink when gambling Regular gamblers drawn from the community University student gamblers

Similarities to Blaszczynski & Nower’s (2002) and Lesieur’s (2001) subtypes (and to those mentioned by Grant, 2007): ENH antisocial, impulsivist gamblers; impulsive action seekers COP emotionally vulnerable gamblers; impulsive escape seekers LOW ER behaviorally-conditioned gamblers; normal gamblers

Researchers Proposed SubtypesStewart et al. Coping Enhancement(Studies 1-4) Gamblers GamblersLesieur Impulsive Impulsive(2001) (2 of 3) Escape Seekers Action SeekersMcCormick Recurrently Chronically (1987) Depressed Under-StimulatedBlaszczynski et al. Depression- Boredom-(1990) Prone ProneBlaszczynski & Nower Emotionally Antisocial(2002) (2 of 3) Vulnerable ImpulsivistGrant Anxious/ Pleasure/Urge &(2007 – yesterday’s Depressed/ General Impulsivity talk) Obsessional Need for Stimulation

Previous Sub-Typing Schemes

Studies 1-4: Conclusions Contrary to hypothesis, no ‘pure coping’ motivated

gambler was identified when coping-motivated gambling occurs, it occurs most often in

combination with enhancement-motivated gambling, although coping appears the primary motivation

Remains to be determined if therapy needs to focus on the primary motive, or both motives given both are elevated

Unexpectedly, a third group emerged which was characterized by low levels of both positive and negative situation gambling More research needs to focus on understanding the

underlying motives for gamblers in the Low Emotion-Regulation cluster; external vs. internal motives (Cooper, 1994)?

Studies 1-4: Conclusions Relations of Gambling Clusters to Motivations for

Drinking (Study 1) At least among problem gamblers who drink when gambling, the

underlying motivations for the two behaviors are similar Consistent with third variable interpretation of high co-morbidity

of alcohol – gambling disorders Possibility of motives as “traits”

Further evidence of utility of sub-types: Vary in expected ways in terms of reward vs. relief gambling

outcome expectancies (Study 2) Varying expectancies may be a function of differing sensitivities

to various reinforcing consequences of gambling (e.g., ENH gamblers more sensitive to reward consequences of alcohol plus gambling; Study 4)

Clinical Implications (Studies 1-4) Results may be useful in designing sub-type specific

interventions for pathological gamblers aimed at level of underlying motives (cf. Conrod et al., 2000b)

Study 1 results may also be useful in helping design interventions for co-morbid problem gamblers – problem drinkers Evidence for consistency of motives across gambling and alcohol

use suggests that the particular motives underlying both forms of addictive behavior should be a primary focus of treatment in a given co-morbid client. Gamblers in the Enhancement gambling cluster should have their

underlying enhancement motives for gambling/alcohol use be a primary target of therapy.

Gamblers in the Coping gambling cluster should have their primary coping motives for gambling/alcohol use be primary targets of therapy (at least).

Clinical Implications Study 2 results may also be useful in helping design

content of matched interventions for gambler subtypes Gamblers in the Enhancement gambling cluster should have their

reward gambling outcome expectancies be a focus of treatment using expectancy challenge techniques (Darkes & Goldman, 1993)

Gamblers in the Coping gambling cluster should have their relief gambling outcome expectancies (at least) be a focus of treatment using expectancy challenge techniques (Darkes & Goldman, 1993)

Need more information on the motivations for gambling in the Low Emotion-Regulation subtype before we can design effective interventions for their gambling; other types of gambling expectancies/other faulty cognitions in this group?

Clinical Implications Study 3 results

Suggest treatment matching may also be useful in prevention / early intervention since similar subtypes evident even among young, relatively less experienced gamblers (see Conrod et al., 2006; Watt et al., 2006)

Study 4 results Suggest the possible utility of naltrexone as a matched

treatment for ENH gamblers (O’Malley, 1996; Peterson et al., 2006) Consistent with suggestions by Grant (talk yesterday)

Proposed Treatment Matching Study for Gamblers

Stewart, Barrett, & Hodgins

(grant under review)

problem gamblers

TAU Tx TAU + ENH Tx TAU + COP Tx

TAU control Mismatched Matched



ENHCOP

Low ER

ENH COP

Low ER

ENH COP

Low ER

ENH

Low ER

COP

pre

post

F.U.

Expected Results 1

0

2

4

6

8

10

12

14

16

control mis-matched

matched

pre

post

follow-up

Expected Results 2

0

2

4

6

8

10

12

14

16

LER ENH COP LER ENH COP LER ENH COP

pre

post

follow-up

TAU + COP TxTAU + ENH Tx TAU ALONE

Gambler Subtype

Questions?

Documents

Sub-typing Gamblers on the Basis of Affective Motivations for Gambling: Implications for Treatment Matching Dr. Sherry H. Stewart, Departments of Psychiatry