Background The Pacific Islands Families (PIF) study has been following a cohort of Pacific children since the year 2000. The purpose of this prospective study is to determine the pathways leading to optimal health, development and social outcomes for Pacific children and their families. Pacific peoples are at high risk for developing problem gambling (the highest risk of the ethnicities living in New Zealand) and have shown heterogeneous differences between the different Pacific cultures in relation to gambling. This highlights the need for significant further study in this area. The longitudinal cohort PIF study has offered a valuable and unique opportunity to study gambling and problem gambling within a Pacific family and child development context, allowing for sub-analyses of the major ethnic Pacific groups and the potential to begin identifying risk and protective factors in the development of problem gambling. In April 2006, the Gambling Research Centre at Auckland University of Technology was commissioned by the Ministry of Health to conduct the research project Problem Gambling - Pacific Islands Families Longitudinal Study. The purpose of this project was to enhance and add value to the existing PIF study by incorporating a substantial gambling component in the six-year data collection phase. Methodology A range of gambling-related questions was incorporated into the interview questionnaire protocols for mothers and fathers of the cohort children at the six-year data collection phase. The questions related to gambling participation and to having problems because of someone else’s gambling, and included problem gambling screens (Problem Gambling Severity Index [PGSI] for mothers and fathers, and South Oaks Gambling Screen - Revised [SOGS-R] for fathers only). All cohort parents (mothers and fathers) were invited to participate in the PIF study six-year assessment. In keeping with previous procedures, all participants were visited in their homes by gender- and ethnically-matched interviewers to complete the structured assessments. Results and discussion This study has significantly increased the knowledge around Pacific peoples’ gambling since the nature of the general population cohort has allowed for analyses to be performed by different Pacific ethnicities and other cultural and demographic variables, which is not usually possible in general population studies due to small Pacific participant sample sizes. Whilst the data in this report represent a cross-section in time, at the six-year data collection point for the cohort, the potential exists for gambling to continue to be a significant part of future data collection phases. This will allow for longitudinal analyses to explore the links between parental gambling and child development of gambling behaviours, as well as risk and protective factors for problem gambling amongst not only adults but also children as they progress through teenage years and into adulthood. It will also allow for exploration of changes over time in regard to gambling participation and problem gambling risk and protective factors. Gambling participation was lower amongst the participants in the cohort than would be expected though a bimodal distribution of gambling (low numbers of people gambling with those who do gamble reporting higher than average expenditure on gambling) was apparent, as was expected from previous national prevalence surveys. Thirty-six percent of all mothers and 30% of all fathers reported that they had gambled in the previous 12 months. Of the mothers and fathers who had gambled, Lotto was the form of gambling most played (89% mothers, 88% fathers) with much lower levels of participation in other forms of gambling. Gender differences were apparent for non-Lotto forms of gambling with mothers participating in Housie and Instant Kiwi gambling (both at 11%) and fathers participating in casino electronic gaming machine (20%), non-casino electronic gaming machine (15%) and Instant Kiwi (14%) gambling. The most preferred forms of gambling were Lotto (80% of gamblers) followed by Housie (9%) for mothers and Lotto (78%) followed by horse/dog race betting (6%) and sports betting at the TAB (5%) for fathers. Amongst those who gambled, four percent of mothers and 16% of fathers were classified as moderate risk or problem gamblers using the PGSI. Using the SOGS-R, 10% of fathers were classified as problem or probable pathological gamblers. Ethnicity appeared to be a key factor in mothers’ gambling but not for fathers. Tongan mothers were less likely to gamble than Samoan mothers; however, those who gambled were 2.4 times more likely to be classified as at risk/problem gamblers, indicating that Tongan mothers are at higher risk for developing problem gambling. Cultural orientation appeared to be related to gambling (in some cases, less gambling) both for mothers and fathers, though different orientations were associated with gambling for the different genders. Fathers who were in the higher total net weekly household income brackets (>$500) were more likely to gamble than fathers in the lower income bracket (<$501), whilst mothers with post-school qualifications were less likely to gamble (0.7 times) than mothers with no formal qualifications. Further gender differences were noted in terms of associations between gambling and health. For fathers both gambling and at risk/problem gambling were associated with psychological distress. Fathers who gambled were more likely to be perpetrators as well as victims of verbal aggression than fathers who did not gamble, with at risk/problem gambling also being associated with physical violence. These findings were not noted amongst mothers whereby at risk/problem gamblers were significantly less likely to perpetrate violence than non-problem gamblers. Not unexpectedly, smoking and alcohol consumption (particularly at higher/harmful levels) were associated with gambling (though not with at risk/problem gambling) both for mothers and fathers. In addition, mothers who drank alcohol were also more likely to have a weekly gambling expenditure in the upper quartile (≥$20) than mothers who did not drink, with increased frequency and amount of consumption associated with increased risk of higher gambling expenditure; this finding was not noted amongst fathers. In addition, a clear association was noted between higher (upper quartile) expenditure on gambling and being classified (PGSI) as a low risk/moderate risk/problem gambler with at risk/problem gambler classified mothers three times more likely, and at risk/problem gambler classified fathers six times more likely to spend in the upper quartile on gambling than non-problem gamblers. The problem gambling screens used (PGSI for mothers and fathers and SOGS-R for fathers only) showed very good internal consistency (reliability). There was good agreement between the PGSI and SOGS-R with 94% of fathers identified as problem gamblers by the SOGS-R also being classified as at risk/problem gamblers by the PGSI. In addition, questions related to lying about gambling and betting more than intended also associated well with the PGSI and SOGS-R within this Pacific cohort. The results suggest that the use of any of these problem gambling screens may be valid for use within a general Pacific population, though this would need to be further tested. Four percent of mothers and ten percent of fathers reported that they had experienced problems because of someone else’s gambling. The findings detailed in this report indicate that different gender and ethnic differences exist amongst Pacific people who should, therefore, not be considered as a homogeneous group. This has implications for service provision by organisations providing services for Pacific people as well as social marketing campaigns around gambling and problem gambling.