ABACUS Counselling Training and Supervision, Welcome to our web site
ABACUS, a New Zealand Qualifications Authority (NZQA) provider of services, a nation-wide provider of Professional Supervision for the workplace.
HomeArticles & PublicationsConferences & PresentationsContact & LocationIntroducing ABACUSSupervisor Login
Gambling Problems and Attempted Suicide: Part I – High Prevalence Amongst Hospital Admissions

International Journal of Mental Health & Addiction
Vol 4(3) August 2006

By Alison Penfold
Simon Hatcher
Sean Sullivan
Nicola Collins


In this study 70 patients admitted to an Auckland (New Zealand) hospital following a suicide attempt were screened with a brief problem gambling screen (the Eight Screen) by hospital staff. Twelve (17.1%) were positives for problem gambling, with half having attempted suicide in the past. A conclusion was drawn that problem gambling amongst those who had attempted suicide could be an important factor in designing effective future interventions.

Key words
Problem gambling, suicide and attempted suicide, screening, Eight Screen


In New Zealand, 1.3% of the general population has been identified as having moderate or severe gambling problems, with Maori and Pacific peoples having a higher risk (Abbott & Volberg 2000).

Suicidal ideation is recognised to commonly accompany problem gambling behaviour (APA 1994). DSM-IV notes:

"Of individuals in treatment for Pathological Gambling, 20% are reported to have attempted suicide" APA p616

The Australian Productivity Commission (1999) identified from available international published research that between 4%-31% of problem gamblers had attempted suicide. If the correct prevalence is within this range, identifying the presence of gambling problems and its influence on decisions to attempt suicide may be important in reducing later attempts. What remains unknown in research to date, is whether problem gambling is a causative or mediating factor in suicide attempts (The Wager July 19th and 27th 2005). This may be more difficult to determine than it may seem as it will largely depend on the attribution and perception of the person attempting suicide. For example, rather than problem gambling being perceived as the cause, instead the attribution may be family break-up, criminal prosecution, creditor pressures, isolation from others, and many other factors that follow from problem gambling (and from resulting depression). Others (MacCallum & Blaszczynski 2003) have concluded that they have been unable to directly link suicide and pathological gambling.

However, identifying whether problem gambling exists amongst those who are at risk for suicide may be a more immediate issue than prevalence rates, especially when an appropriate intervention may prevent a further, completed suicide attempt. Recent research indicates that non-fatal attempts are one of the strongest predictors of an eventual fatal attempt (Hirschfield & Russell 1997; Lewis, Hawton & Jones 1997).

An aim of this research was to determine the prevalence of problem gambling behaviour amongst those who had recently attempted suicide, as a first step towards providing an appropriate intervention to reduce future suicide risk.


The study was conducted in the Accident and Emergency Department of an Auckland hospital. Approximately 450 people present to the hospital each year following an attempted suicide/self-harm episode, from a hospital area catchment of 445,000 people.

All patients who attended the hospital for a 20-week period, following a suicide attempt/self harm episode, were to invited to complete a staff administered questionnaire, including a brief problem gambling screen. Demographic information was accessed from either the patient or from records.

The Eight gambling screen (an acronym for Early Intervention Gambling Health Test –copy annexed) was originally developed for use by family doctors and other health workers in primary care environments (Sullivan 1999). It is usually self-administered, however a version developed for phone counselling was used in this project where the hospital staff would administer the screen verbally to patients. Despite its brevity (eight questions covering emotional, cognitive and behavioural dimensions) it has been found to have a high sensitivity for Pathological Gambling Disorder in treatment and forensic environments. A positive answer to four or more of the eight questions indicates the existence of a gambling problem.

Additional screens were included in the questionnaire but are not reported on here. Patients were excluded from the study if they were psychotic or were already registered as a patient under the Mental Health Act (1999).


During the 20 week period that data was collected, 189 patients were admitted to the hospital following 203 suicide/self harm attempts (some attempted harm more than once during the period). Female patients comprised 74%, and males comprised 26% of the sample.

The ethnicity of the patients reflected the Maori and NZ European composition of the Auckland region where the study took place, however Asian and Pacific people were under-represented amongst these patients while other ethnicities were over-represented.

Table 1: Ethnicity of all patients (n=189) admitted to the hospital during the period following suicide/self harm attempts compared with ethnicity of Auckland population* (percentages were rounded up and may exceed 100%)
  Maori NZ European Pacific Asian Indian Other
Patients 12%(22) 64%(121) 2%(3) 2%(3) 3%(5) 19%(35)
Auckland 9% 69% 11% 11% 1%
*source Statistics NZ

Past and current attempts
42% of these patients had attempted suicide/self harmed at least once previously. The most common means of harm used was an overdose of a drug, with 75% using this means solely and a further 11% using drugs in conjunction with another method.

Participation in this study
Of these 189 eligible patients, 70 patients participated in the further data collection process. The data collection process was confined to the period following admission up to discharge from the hospital.

The predominant reason for patients (n=119) who were eligible not participating, was that the admission was during the evening, when limited numbers of staff were available to conduct the interview.

Table 2: Reasons for exclusion or non-participation (n=119)
Reason Too confused
to ask
Denied intention
to harm self
Too young Not asked Refused
%(n) 4%(5) 3% (4) 1% (1) 80% (95) 12% (14)

The questionnaire was read to the patient as part of the hospital’s usual admission process and responses were recorded. Demographic data was added later from routinely collected admission information, and included age, gender, living arrangements, and details of past attempts. Details of alcohol use are reported in the second paper of this project.

Gender and age of participating patients
Two-thirds (64%) of participating patients (n=70) were female, and one-third (36%) male, a similar gender distribution to all patients admitted during the period following an attempted suicide/self harm incident.
The mean age of participating patients was 32 years (s.d. 13 years), similar to all patients (31 years; s.d. 12 years) (t test; p=0.25).

Ethnicity of participating patients
The ethnicity of participating patients (n=70) was similar to all patients (n=189).

Table 3: Ethnicity of participating patients (n=70) admitted to the hospital during the period following suicide/self harm attempts.
  Maori NZ European Pacific Asian Indian Other
%(n) 17%(12) 57%(40) 4%(3) 1%(1) 3%(2) 17%(12)

Past and current attempts of participating patients
39% of participating patients had attempted suicide/self harmed more than once in the past, similar to all patients (41%) and using similar methods (drug overdose (74%) or drug overdose in conjunction with another method (10%).

Problem Gambling
Twelve (17%) of the participating patients (n=70) scored as positive (4+) using the EIGHT gambling screen (95% CI; 9.2%-28% A further nine patients scored 1-3 on the gambling screen.

Ethnicity of problem gambling patients

Table 4: Ethnicity of participating patients (n=70) admitted to the hospital during the period following suicide/self harm attempts compared with ethnicity of patients identified as problem gambling (percentages were rounded up and may exceed 100%)
  Maori NZ European Pacific Asian Indian Other
Participating Patients
17%(12) 57%(40) 4%(3) 1%(1) 3%(2) 17%(12)
Problem gambling patients
% of ethnic group (n)
42% (5) 13% (5) 33% (1) 0% 0% 8% (1)

Although there were larger percentages of Maori and Pacific participating patients identified as problem gambling, these were not statistically significant (logistic regression p=.09).

Gambling screen scores
A score of four or more on the EIGHT screen identified problem gambling, with a maximum score of eight possible. Scores of six or more are strongly correlated with Pathological Gambling Disorder. Three-quarters (n=9) of screen positives scored six or more on the Eight Screen, with 50% of the 12 screen positives scoring seven.

Gender of screen positives
Seven were female (16% of participating female patients) and five were male (20% of participating male patients).

Age of screen positives
Age of those scoring positive on the gambling screen (30 years, s.d. 7 years) was similar to all patients and participating patients (t test; p=0.25).

Past and current attempts
50% of those scoring positive on the gambling screen had attempted suicide/self harmed more than once in the past, with 91% using drug overdose, and 8% using drug overdose in conjunction with another method.

Past psychiatric history
58% of participants who scored as positive on the 8 screen had previously been treated for psychiatric problems, compared with 60% of those participants who did not score positive.

Participants compared with non-participants
There were no statistically significant differences between those who participated and those who did not, in respect of their age, gender, ethnicity, past psychiatric history, and past suicide attempts.


The proportion of those admitted to the Accident and Emergency Department of the hospital following a suicide attempt/self harm episode was high (17%) compared with estimates of problem gambling in the general New Zealand population (1.3%; Abbott et al). Patients who are admitted to hospitals following a suicide attempt might be expected to have a range and severity level of health problems that exceed those found in the general population (i.e. there may be a selection bias). In addition, depression is a strong indicator of attempted suicide (Newman & Thompson 2003) and may be a confounder in this study population in that over half (58%) of those who scored positive for problem gambling had a past history of psychiatric problems, including depression, which may also have pre-dated their gambling problems. This is partly balanced by equal numbers of participants without gambling problems (60%) also having had a past psychiatric history. However, the participants may comprise a problem group more severely affected with health problems than the general community. Therefore we cannot posit that problem gambling may be a strong factor in suicide, but the finding that problem gambling does exist for one in six who attempt suicide, is an important clinical fact for treatment.

A substantial proportion of the patients were not invited to participate in the study due to the condition of the patient or for other reasons, although this did not appear to result in any systematic errors. Analysis of possible confounding factors (e.g. age, ethnicity, and past psychiatric history) identified that the participant and non-participant group were not statistically different.

The numbers of participants in the study were not as high as desired for the reasons above, while the numbers screening positive were also low in number, but surprising in percentage. Nevertheless, this was research that enlisted from a different population than usual, namely those seeking help for gambling problems. This population comprised those who had been hospitalised immediately following a suicide attempt, and involved participation of hospital personnel who are required to care for patients following a life-threatening event. Problem gambling can appear to be low on a hierarchy of the patient’s needs and recruiting psychiatrists and other specialist hospital staff to screen for a behavioural addiction may mitigate against problem gambling research with this population. This project was able to overcome these barriers, and contributes information to the body of knowledge of problem gambling from an unusual perspective.

Half of the problem gambling positives had attempted suicide in the past indicating that these patients were at high risk for future attempts. Gambling problems are often not transparent, especially where the suicide has been completed. Depression may be a stronger driver for suicide, but if it followed the problem gambling, then depression in turn may be driven by the gambling. In addition, even if depression preceded the gambling, stressors arising from the gambling may intensify the depression and the likelihood of suicidal ideation. The findings of this research indicate that further information may be required as to the influence of problem gambling in the decision to attempt suicide and may assist to identify appropriate interventions to reduce future attempts.


DSMIV (1994) refers to the high risk of suicide for those meeting Pathological Gambling Disorder. However, the true level of rates for suicide attempts remains unclear because populations of problem gamblers in treatment surveyed, are often not able to be generalised to all problem gamblers. However, knowledge of risk amongst these special populations provides important information and directions for addressing their specific needs, whether or not they clarify the rate of risk for all problem gamblers. This research identified that more than one in six who attempted suicide were experiencing problems from gambling, and that from their high scores, the majority were experiencing a range of effects that were likely to meet Pathological Gambling Disorder criteria. Further research is required to determine the effect that their gambling may have on future risk for suicide, however the offer of counselling for these gambling problems may be an appropriate interim intervention pending such research.


Abbott M & Volberg R (2000) Taking the pulse on gambling and problem gambling: a report on Phase One of the National Prevalence Survey. Wellington, Department of Internal Affairs.

American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders. 4th Ed. Washington DC, APA.

Australian Productivity Commission (1999) Australia’s Gambling Industries. Canberra, APC.

Blaszczynski A & Farrell E (1998) A case series of 44 completed gambling related suicides. Journal of Gambling Studies 14, 93-109.

Hirschfield R & Russell J (1997) Assessment and treatment of suicidal patients. New England Journal of Medicine, 337, 915-919.

Lewis G, Hawton K & Jones P (1997) Strategies for preventing suicide. British Journal of Psychiatry, 171, 351-354.

MacCallum F & Blaszczynski A (2003) Pathological gambling and suicidality: an analysis of severity and lethality. Suicide and Life Threatening Behavior, 33 (1), 88.

Sullivan S, Abbott M, McAvoy B & Arroll B (1994) Pathological gamblers –will they use a new telephone hotline? New Zealand Medical Journal 107, 313-315.

Sullivan S (1999) The GP ‘Eight’ Gambling Screen. Doctorate, Department of General Practice, University of Auckland, Auckland.

The Wager (2005) Suicide and pathological gambling – the state of the evidence and need to improve scientific methods (Parts 1 & 2) Volume 10 (8 & 9) http://www.basisonline.org/wager/