The concerned significant others of people with gambling problems in a national representative sample in Sweden – a 1 year follow-up study

Problem gambling is regarded as a public health issue [1–3] and is often defined as “gambling behaviour that creates negative consequences for the gambler, others in his or her social network, or for the community” [4]. Previous studies have shown that problem gamblers suffer higher levels of health-related problems than the general population [5–7]. Problem gamblers report lower levels of general and mental health than others, and suffer from higher rates of depression, anxiety and suicide ideation than the general population [8–14].

Concerned significant others (CSOs) is a term used in earlier research [15] and will be employed in this paper to describe people in the surrounding environment of a problem gambler. In this paper, a CSO can be a parent, spouse, child, relative but also a friend or colleague. This definition is coherent with Hodgins et al. [15] while it is broader than the definition of Wenzel et al. [16] who examined only family members (15-16) and Ingle et al [17] who included spouses and boyfriends/girlfriends. Every problem gambler is estimated to have between 7-16 CSOs who could be affected by his or her gambling [18–20]; the range depends on how CSOs are defined. The Australian Productivity Commission [20] reported that almost half of all problem gamblers (49.4%) live in households with an average of two children, while their own survey estimated that there are statistically 0.6 children (under the age of 15 years) living with the average problem gambler. Studies have also examined and confirmed the association between being a problem gambler and having a parent with a problem gambling history [21, 22]. However, little attention has been directed to the negative consequences of problem gambling on personal and social relationships [23, 24], and the empirical research on consequences for close relatives and friends is surprisingly scarce [25, 26]. There is extensive research to suggest that consequences for CSOs range from economic problems to social and health problems [27], however, it has used help-seeking populations or convenient samples to explore the situation of CSOs [17].

Most of the research to date on the impact of problem gambling on CSOs has focused on the female spouses of problem gamblers [16, 23, 28–30]. Spouses have reported harassment and legal threats by the gambler’s creditors, distress in their children, additional responsibilities arising from the gambler’s absences and neglect of family and physical manifestations such as headaches, sleeping difficulties, depression and anxiety [31–33]. Suicide attempts by spouses and partners of problem gamblers are reported to be three times that of the general population [34, 35]. The National Gambling Impact Study [36] estimated the lifetime divorce rates for problem and pathological gamblers were 39.5% and 53.5% respectively; the same rate for non-gamblers was 18.2%.

Research on help-seeking populations has found that spouses of problem gamblers are exposed to higher levels of domestic violence than those with a non-gambling partner. Korman et al. [37] showed that 25.4% of problem gamblers reported that they perpetrated severe intimate partner violence. This result is consistent with findings from a previous study [38], and from earlier research on spouses [32, 39] and perpetrators [40]. Children of problem gamblers have been reported to be at risk of experiencing abuse by both the gambler and his or her spouse [41]. Lesieur and Rothschild [42] suggested that children of problem gambling parents are two to three times more likely to be abused than their peers. However, Schluter et al. [43] found no association between being a relative of a problem gambler and being exposed to domestic violence in a longitudinal research study on South Pacific families in New Zealand.

The understanding and treatment of clinical problem gambling is crucial for supporting the concerned individual and his or her family. However, a public health perspective addresses not only the risk of problems for the individual gambler but also the quality of life of the families and communities affected by gambling and acts to prevent the problems at a more structural level [1]. From a public health perspective, efforts targeted at an individual level are often of limited value for society as a whole [44]. We are not able to make more general assumptions from non-representative samples so therefore, knowledge based on epidemiological data is important.

One identified study has explored the epidemiology of problem gambling in the family using a national representative [16]. This Norwegian study showed that problem gambling has a strong impact on the quality of life of CSOs and that 2.0% of the population was a CSO of a problem gambler. Most of the CSOs were women, urban dwellers, divorced, and had unsatisfactory financial situations and unsatisfactory subjective health. The effects of gambling on the CSOs were worsening family financial situations, decreased social relationships, and increased conflicts levels within the family. They also reported greater mental health problems than in the general population, including sleep disorders, depression symptoms, obsessive-compulsive symptoms, alcohol problems, substance abuse and their own problem gambling [16]. Exposure to violence was not included as an outcome in the study. However, the study had a low response rate (36.1%) and in addition, it did not include any separate analyses for men and women.

The proportion of problem gamblers in Sweden seems consistent at just over 2% of the total of the population: 3.2% of the male population and 1.3% of the female population. The highest proportion of problem gambling is found amongst men aged 18-24 years old with close to 1 in every 10 men having a gambling problem. In the oldest age group (65-84 years old) the proportion of problem gambling is higher among women than men [45]. As reported in the first section of this paper; problem gamblers are exposed to various problem regarding health, finances and social relations. Close to 4% of the population (n = 262,500) live in the same household as a problem gambler. Further, close to one third of problem gamblers live in a household that includes children [45].

The hypotheses in this study are based on the findings of Wenzel el al. [16] that CSOs were more likely to be women, were associated with problem gambling, riskier alcohol consumption, and experienced more mental health problems. Additional research questions are included such as social relations, financial situation and important life events from a longitudinal perspective.

Data was derived from a Swedish longitudinal survey on gambling and health (Swedish Longitudinal Gambling Study - Swelogs). The study used data from the two first measurement points, the baseline study, Wave I and the 1-year follow-up study, Wave II. Both Wave I and II used the same methods of data collection. The primary method was computer-based telephone interviews conducted by Statistics Sweden. Mail surveys and reminders were sent to non-responses.

Wave I used a stratified simple random sample based on gender, age group and estimated risk of problem gambling (being male, low income earner, or living with family members on social welfare) of 15,000 participants. The estimated risk was calculated from register variables according to results from a pilot study conducted in spring 2008.

The frame population consisted of 7,320,367 persons aged between 16 and 84 years in the Swedish Total Population Register. The stratified sample was drawn from the frame population. Calibration weights were used to adjust the results to the frame population.

The total number of respondents (n = 8,165) corresponded to a weighted response rate of 63%. The internal attrition for the question regarding being a CSO was 3.3% (n = 264).

Wave II was conducted from December 2009 to March 2010. The 8,165 individuals who were respondents in Wave I were re-contacted and new calibration weights were calculated. The weighted response rate for the sample in Wave II was 74% (n = 6,021). There was no significant difference between the attrition of CSOs and the general population. In addition, we found no differences in the attrition when comparing male and female CSOs and the general population regarding age, country of birth or family situation. Male CSOs who only participated in Wave I were slightly more likely to be on social welfare than other men while female CSOs who only participated in Wave I were significant more likely to have a lower level of education compared with women in the general population.

In the total sample, 18.2% (n =1,472) defined as a CSO. Men were slightly more likely to be a CSO at 19.5% (n = 778) compared with 17.5% (n = 694) for women (OR 1.2; p < 0.01). This result dismisses the hypothesis that women are more likely to be CSOs. As Table 3 shows, few people aged 65 years old and older or people with a higher level of education were CSOs. Being an immigrant was associated with being a CSO for men only. Further, for men being a CSO was most prevalent among singles with and without children, and married men with children. The last group was only significant after being controlled for age. The family situation for women was less clear, even if single mothers were over represented and married women or women who lived together with a partner without children were under represented. When age and problem gambling was controlled for, only being a single mother was a significant factor.

All CSOs had lent money to someone who they thought or knew would use it to gamble or pay gambling debts. Approximately one third of men who were CSOs had done this compared with 13% of the women who were CSOs. However, the OR was higher for women than for men. The explanation for this is found in the fact that men who were not CSOs also (to a relatively high degree) had lent money to someone for gambling purpose (6.3%). A low proportion of both male and female CSOs had sought help, 8.5% of male CSOs (n = 66) and 10.9% female (n = 75); thus, there was no significant difference between sexes (not shown in table).

In contrast to the hypothesis that women were more likely than men to be CSOs, this study showed that men are equally likely or even more likely, to be CSOs. Further, men essentially face the same difficulties as female CSOs. However, Swedish statistics from a problem gambler helpline show that women represent the majority of close relatives of problem gamblers seeking help, 80% of annual calls from close relatives are from women [51]. Only a small number (approximately 10%) of the CSOs in this study had sought help which indicates that various services should improve their accessibility and knowledge on issues for CSOs of problem gamblers. Earlier research has shown that the involvement of significant others in treatment is helpful for the problem gambler [52, 53], a study of the impact of significant others (e.g. spouses and partners) to 4,411 problem gamblers who were discharge from treatment found that having a significant other was associated with the odds of effective treatment [17].

The hypothesis that CSOs were more likely to be associated with problem gambling, riskier alcohol consumption and mental health problem was partially supported. Prevalence rates from Wave I show that all CSOs experienced mental health related problems and had riskier alcohol consumption than the general population. However, only men who were CSOs were associated with own problem gambling.

As stated, while Wave I showed similarities between male and female CSOs, important gender differences were found. As noted above, men who were CSOs were more likely to be problem gamblers than men in the general population and than women in general including women who were defined as CSOs. Female CSOs reported a worse general health situation in terms of self-reported general health and sick leave days as well as months of absence from work because of illness while male CSOs reported more fear of losing employments. Male CSOs were less absent from work because of illness than other men. Male CSOs also reported more legal problems both at Wave I and II, and male ex-CSOs had fewer legal problems than men who were CSOs at both time points. Female CSOs reported a lack of social support. It is clear from these gender differences that male CSOs, consistent with studies on male problem gamblers, experience more work, debt and legal problems. In contrast, female CSOs, like female problem gamblers, tend to experience problems with relationships and have more physical and mental difficulties [46]. However, in the analyses of CSOs at Wave II, women defined as CSOs at both time points were more likely to have problems at work. This was not found for men who were defined as CSOs at both time points.

Our findings regarding experiences of violence confirmed the results of Brasfield [40], Muelleman et al [39], Korman et al [37] and Bland et al [38]. In contrast, Schluter et al. [43] did not find any association between being a relative of a problem gambler and exposure to domestic violence. However, the authors suggested this could be because of the study design. Both male and female CSOs experienced more violence than the general population. As we do not know the nature of the relationships between the CSOs and the person with gambling problems, we do not know the context of the violence.

Logically, financial hardship, such as in finding it hard to pay the bills or having to receive social welfare, was associated with being a CSO. Both male and female CSOs were more likely to encounter financial hardship at both time points. Women who defined as ex-CSOs in Wave II reported fewer difficulties in paying bills in Wave II compared with women still defined as CSOs. Both male and female CSOs had lent money to others for gambling purposes. This behaviour was more common among men while the OR was higher for women. The explanation for this is found in the fact that other men in the general population, as opposed to women in the general population, (and to a relatively high degree) had also lent money to someone for gambling purpose or to pay gambling debts.

Even though some studies have shown that women who were married to problem gamblers tended to have long marriages [47], earlier research has shown that problem gambling can damage family relations [16]. The relationship between the problem gambler and the CSO in our study is unknown. However, all CSOs had more arguments with someone close to them and were more likely to have been divorced or separated between Wave I and II than men and women in the general population.

Our results are in line with the findings of Wenzel et al. [16], even if our separate analyses for men and women showed that it was only male CSOs that were more likely to be problem gamblers. This and other gender differences in our results suggest that it is important to look at women and men separately. There are no studies to date that examine CSOs from a gender perspective despite the fact that being a CSO is often interpreted as being female and that samples have been dominated by female partners. The use of combined samples of men and women and the analysis of such samples as one entity or purely female samples can lead to gender effects being ignored or missed. The number of female problem gamblers has been increasing in some parts of the world, including Manitoba, Canada and Australia [54, 55]. Thus, further research is needed into the male spouses of female gamblers [21, 51], same sex spouses and other types of relationships.

Wenzel et al. [16] found a prevalence rate of 2%; however this study found that 18.2% of the population were CSOs. The Norwegian study had a narrower approach asking each respondent two questions (based on the lie-bet instrument) regarding their close relative. In contrast, this study asked a wider and more open question - did the respondent have someone close to them who has, or has previously had, problems with gambling, of which 18% of the sample answered yes to. Further important findings include the long-lasting effects of the associations between being a CSO and mental health, social relations, and financial hardship. In addition, there seems to be opportunities to change when CSOs - for reasons unknown to us – become ex-CSOs. Wenzel et al. [16] did not control their results for problem gambling or age. Our study only found an association between problem gambling and male CSOs, which made it important to consider in analyses. When controlling for problem gambling, we found a tendency for the OR for men to decrease and increase for women.

The results of this study should be carefully interpreted in light of the study’s limitations. One central limitation of the study is that we do not know the nature of the relationship between the CSO and the person who was reported to have gambling problems. Since we asked the respondents not only if they had someone close who only has gambling problems but also if this person had previously had gambling problems, it is possible that the period of gambling problems was a long time ago. Of note is that the person close to the respondent who was reported as having problems with gambling is not defined as a problem gambler because it is the respondent’s own perception of gambling problem that is reported. Even if we do not know the nature of the relationships, they are not likely to be only close family as 18% of the population were defined as CSOs and we know that 4% of the population lives in the same household as a problem gambler. However, even if being a problem gambler as defined by the PGSI 3+ is not comparable with having some kind of gambling problem, the difference between 4 and 18% indicates that the CSOs in this paper also include friends and colleagues. And the whole group is more likely to face various problems. The causal relationship between being a CSO and the problems is however, unknown.

In addition, the differences between male and female CSOs may be due to the fact that men in general have “looser” relations with the individuals with gambling problems (as friends, colleagues, and more distant family members) while for female CSOs, this person could possibly be their partner or close family. There is, as presented in the introduction, research performed on relatives of problem gamblers. However this research has been nearly exclusively clinical. What kind of support is needed for friends or colleagues has not been on the agenda, as well as that more epidemiological data on relatives as well as CSOs (as we define CSOs).

The strengths of the study are that it points out areas of interest, that CSOs are a broad and important group and that the associations between various problems are probably not only found in the close environment but also in the wider social network. A further strength is the longitudinal approach that enables to see changes over a one-year period. However, this design was not used in this study to examine the causal relationships between being a CSO and social, financial and health factors, but we now know how these factors developed for the group over a period of one year. A possible way forward in future studies is to examine different groups of CSOs from a gender perspective. CSOs are important, both for reaching and helping people with gambling problems and for their own sake.

Jessika Svensson, PhD in Public Health at the Mid Sweden University. She works partly at the Swedish National Institute of Public Health and primarily works on the Swedish Longitudinal Gambling Study. Contact: Jessika Svensson, Swedish National Institute of Public Health, SE 831 40 Östersund.

Ulla Romild, PhD in Statistics, holds a position as Public Health Planning Officer at the Swedish National Institute of Public Health. She is primarily involved with the Swedish Longitudinal Gambling Study, SWELOGS where she is responsible for the data analysis and also bears the scientific responsibility for the project. Ulla also holds a part time position as Statistical Consultant at The Department of Research and Development (RaD), Levanger Hospital, Health Trust Nord-Trøndelag, Norway.

Emma Shepherdson Public Health Planning Officer at the Swedish National Institute of Public Health.