Gambling and its harmful effects on human health and well-being are increasingly recognized as significant public health and policy issues in many countries [1]. While there is no clear evidence on the directionality, there is interrelation between excessive gambling and multiple negative effects on human health and health behaviors, such as mental health difficulties and substance use [2, 3]. Moreover, gambling is also linked to variety of social harms, including financial hardship, unemployment, and relationship problems [4].

Of the various gambling products, electronic gambling machines (EGMs) have been shown to be among the most harmful forms of gambling, mainly due to the structural features of EGMs, such as speed, near misses and high event frequency combined with attractive audiovisual elements [5, 6]. In addition to structural characteristics, a significant concern is the consistent finding across jurisdictions in Europe, North America and Oceania that the accessibility and availability of EGMs are greater in socioeconomically more disadvantaged neighborhoods than in advantaged ones [7,8,9,10,11,12,13,14,15]. Neighborhood accessibility and the availability of EGMs have been associated with increased gambling participation, higher rates of gambling harm [5, 16,17,18,19] and higher rates of seeking help for problematic gambling [20]. Exposure to gambling is linked to higher expenditure rates [8, 14, 21], which in turn predict harm [22].

In gambling research, availability and accessibility are often used interchangeably to refer to different dimensions of physical exposure to gambling products [23]. Availability is typically measured as the number of terrestrial gambling venues (e.g., casinos, arcades) or as the number of EGMs in a given region [21], while geographical accessibility is understood to be the cost of reaching (e.g., distance, travel time, monetary cost) gambling opportunities [5, 23]. Definitions of accessibility, however, vary with accessibility metrics, which sometimes incorporate several measures, such as density, travel cost, attractiveness, and other possible attributes in the models [24, 25]. The focus on density and distance in gambling research aligns with policy relevance, as both can be regulated to prevent harm. This emphasis on policy-relevant factors is consistent with the criteria outlined in accessibility research [26]. Incorporating both density and proximity to spatial analyses of exposure to gambling opportunities is recommended in gambling research [23].

Several studies show that the accessibility of EGMs is higher in socioeconomically disadvantaged neighborhoods when measured by the distance to the nearest venue [8, 12, 27]. For example, a Canadian study revealed a significant negative correlation between the average walking distance to the nearest EGM venue and both average household income (r = -0.378) and the proportion of individuals aged 20 years or older without a high school diploma (r = -0.307) [27]. A Finnish study revealed that for every 1000 euro increase in median income, there was a 0.06 unit decrease in EGM density [11]. Similarly, numerous studies across various jurisdictions have confirmed higher EGM density in socioeconomically disadvantaged neighborhoods [9, 10, 13, 14, 28], which can lead to greater gambling harm for residents.

Local gambling exposure and neighborhood socioeconomic status have been linked to increased average individual expenditures [8, 14, 21]. An Australian study using venue-level revenue data revealed a 0.5% increase in expenditure per adult for every one-point increase in an index of socioeconomic deprivation, with 40% of this effect attributed to EGM density [8]. Likewise, Grumstrup and Nichols [14], utilizing venue-level revenue data in Illinois, United States, reported that a 1 percentage point increase in the poverty rate corresponded to a 1.47% increase in EGM expenditure per capita and a 1.17% increase in EGM density. Using self-reported expenditure data in Canberra, Australia, Marshall et al. [21] found that individuals living within two kilometers of their regularly visited venue had the highest annual expenditure levels.

Research indicates that local area disadvantage and other contextual factors are likely to have an impact on the strength of the relationship between neighborhood disadvantage, accessibility to EGMs, and expenditure on EGMs [29]. For instance, an Australian study found no consistent spatial correlation between gambling expenditure and local level socioeconomic disadvantage, highlighting how contextual factors can influence such relationships [30]. Gambling expenditures in neighborhoods with gambling venues can vary for many reasons: there can be a small group of players in the neighborhood with elevated levels of spending, or the players visiting the neighborhood may spend on gambling [29, 30]. Whether the higher levels of EGM spending stem from local residents or from people visiting the venues has been examined in studies on the catchments of the venues [18, 21, 29]. Marshall et al. [21] reported that there was considerable variation in the sizes of the catchment areas of different EGM venues. Young et al. [18] reported that one-third of customers visited the closest gambling venue, but individuals with an increased risk of gambling harm were more likely to visit the venues closer to their homes. Furthermore, there is evidence that different venues attract different parts of the population [25].

In this study, we contribute to the existing research in the field by analyzing the associations between local SES, geographic accessibility to EGMs, and actual expenditure on EGMs. Prior research indicates that relatively little is still known about whether higher accessibility to EGMs in a neighborhood is associated with a higher level of expenditure for players residing in the neighborhood. From a methodological viewpoint, data on average expenditure are largely derived from self-reports or venue-level revenue statistics. A typical weakness associated with survey data is subjective bias, while venue-level statistics do not reveal the exact residential locations of the players. Overall, previous studies utilizing player account-based expenditure data [31] or grid data on socioeconomic status are scarce [8, 9]. We are unaware of other studies that have combined actual individual-level interaction data between players and EGMs (stakes, losses), the exact residential addresses of players, and high spatial resolution grid data on socioeconomic status. By utilizing these data sources, we seek to significantly advance the current understanding of the associations among the geographic accessibility of EGMs, expenditure, and socioeconomic status at the local area level.

To summarize, the aims of this study are (1) to examine the association between EGM accessibility and the socioeconomic status of different neighborhoods and (2) to examine whether the EGM expenditure of neighborhood residents is associated with EGM accessibility and neighborhood socioeconomic status.