How Does Gambling Correlate with Chronic Health Conditions?

Abstract

This study investigates the relationship between gambling behaviors and health outcomes, focusing on chronic illnesses (hypertension, diabetes).

Using data from the 2023 California Health Interview Survey (CHIS), a mixed-mode survey with address-based sampling, we analyzed a large, diverse sample of California residents. We conducted logistic regression models that assessed the correlation between gambling and health outcomes, adjusting for gender and age.

Our results indicate that individuals who gambled within the past year had significantly higher odds of reporting hypertension (OR=1.35) 95% CI [1.24, 1.48] and diabetes (OR=1.32) 95% CI [1.16, 1.49] in comparison to non-gamblers. Gender and age were identified as significant predictors of health outcomes across all models, while gambling did not show an interaction effect with gender.

These findings highlight the complicated relationships between gambling behaviors and chronic physical illnesses. It is crucial we emphasize gambling as a serious public health concern. This study contributes to the existing literature by using recent and newly available CHIS 2023 data to address gaps in understanding the health risks associated with gambling. Understanding these relationships can help develop future research, interventions, and policies to reduce the negative health impacts of gambling, particularly among vulnerable populations.

Introduction

As of December 2024, gambling is legalized at the state level, thus the legality of gambling varies drastically state by state. Currently, 44 states permit land-based casino operations, and 38 states (plus Washington, D.C.) allow live, legalized gambling. While gambling is legal in Missouri, it has not yet been implemented to its fullest extent.^1,2 According to the American Gaming Association, as of 2023, an estimated 49% of all adults in the US participated in some form of gambling, while 26% visited a casino during the same period, marking the highest gambling participation rate on record in the US.^3,4 Gambling has become increasingly prevalent in the US, particularly with the expansion of online gambling and state-sponsored lotteries in recent years. While gambling can be a recreational activity for some, it may also lead to significant health consequences, including chronic illnesses and mental health conditions.

The high rate of US gambling behaviors is alarming, as pathological gambling has been found to be associated with elevated heart rate and increased cortisol levels, both physiological markers of stress.⁵

Prolonged exposure to elevated cortisol levels is extremely problematic as it contributes to poor health outcomes, including an increased risk of cardiovascular diseases like hypertension and atherosclerosis, as well as chronic conditions such as obesity, hyperglycemia, and insulin resistance—which are all indicative of type 2 diabetes.^6,7

Given the intersection of biological, psychological, and social responses to gambling, our research study and analysis are guided by the biopsychosocial model of health. The biopsychosocial model emphasizes that, as a high-stress behavior, gambling significantly impacts both physical and mental well-being. This study seeks to answer the research question: ‘How do gambling behaviors correlate with physical health outcomes (e.g., chronic illnesses, self-reported health status)?’ We hypothesize that individuals who gamble will exhibit a higher prevalence of chronic illnesses.

The California Health Interview Survey, 2023 (CHIS 2023) provides a valuable dataset for exploring these connections, as it includes data for 21,671 adults.⁹ There are strong links between gambling and both diabetes and hypertension. This study aims to fill a gap in scientific literature by focusing on a large, diverse sample of California residents from the CHIS 2023 dataset. This is particularly significant because the data is both recent and newly included in CHIS 2023, addressing a topic where much of the existing peer-reviewed research dates back to the early 2000s. Given these implications, gambling in the United States is an immensely pressing public health concern that must not be overlooked.

Methods

The data for this study come from the CHIS 2023. This survey is a mixed-mode online and telephone-based survey that aims to collect data on a wide range of topics, including health conditions and general demographic information. For this survey, data is collected via address-based sampling, where one participating household is randomly selected from various geographic regions to obtain a sample that is representative of the state of California. Once a household is selected, one adult is randomly chosen to complete the survey. All households which agreed to provide information are included in the dataset. Open access to this dataset was provided by the University of California Los Angeles and accessed via the website at https://healthpolicy.ucla.edu/our-work/california-health-interview-survey-chis/access-chis-data. The robust sampling method and size of this dataset provides a large, diverse dataset to analyze gambling behaviors and their correlation with health outcomes. The data for this particular study consists of responses solely from 2023 due to it being the first year that questions regarding gambling behaviors were included in the questionnaire.

For the purpose of this study, we focused on binary variables for gambling, hypertension, and diabetes. Gambling behaviors were assessed by a self-reported ‘Yes’ or ‘No’ answer to the question “Have you gambled in the last two months?” Next, we converted the responses into a binary 0 or 1, with yes = 1 and no = 0 for our logistic regression. Furthermore, we evaluated the presence of hypertension by looking at the responses to the question “Has a doctor ever told you that you have high blood pressure?” This question had three possible responses: yes, no, and borderline hypertension. For the purpose of the study, we collapsed the answers yes and borderline hypertension categories into one and made hypertension or borderline hypertension=1 and no = 0 for the logistical regression analysis. Finally, we assessed the presence of diabetes by looking at the responses to the question “Has a doctor ever told you that you have diabetes (non-gestational)? Once again, the responses were either yes or no which we converted into yes = 1 and no = 0 for logistical regression analysis. Finally, we conducted logistic regression to assess the relationship between gambling and health outcomes, adjusting for gender and age. Interaction terms tested whether gambling effects varied by gender. Each model included a crude association, adjustments for gender, age, smoking, binge drinking and we tested for an interaction by age and gender. Results were evaluated at a 95% confidence level.

To control for potential confounders in the relationship between gambling and health outcomes, binge drinking, educational attainment, and health insurance status were operationalized based on responses from the CHIS 2023 survey. Binge drinking was defined using the standard public health definition provided by the CDC, with participants categorized as binge drinkers if they reported consuming five or more drinks on one occasion for men (or four or more for women) at least once in the past 30 days. Responses were coded into a binary variable, where 1 indicated binge drinking and 0 indicated no binge drinking during this timeframe. Educational attainment was operationalized by dichotomizing participants based on their highest level of education completed; individuals who reported having a college degree or higher were coded as 1, while those with less education were coded as 0. Health insurance status was determined by participants’ responses to whether they were currently covered by any form of health insurance, including employer-sponsored plans, Medicaid, or private insurance. Those with any type of coverage were categorized as insured (coded as 1), while those without any form of coverage were classified as uninsured (coded as 0). These operational definitions were chosen to align with established public health metrics and facilitate meaningful comparisons across subpopulations.

Smoking, binge drinking, and education were included as control variables due to their potential to confound the relationship between gambling and health outcomes like diabetes and hypertension. Smoking and binge drinking are established behavioral risk factors for diabetes and hypertension and frequently co-occur with gambling. As a proxy for socioeconomic status, education influences gambling behaviors and health disparities. By controlling for these variables, we aimed to isolate the independent effect of gambling on diabetes and hypertension.

Results

When analyzing our results, we must first look at the sociodemographic characteristics of the study population. In terms of gender, the study population consisted of more females than males with females making up 57.2% of the study population while males made up 42.8%. Despite the study population consisting of more females than males, the proportion of females who self-reported gambling is fairly consistent with the proportion of males who self-reported gambling with 24.9% of females reporting they had gambled in the last 12 months and 28.8% of males reporting gambling as well.

When looking at the racial and ethnic demographics of the study, we see that the study population is 49.5% white, with Asian and Latino individuals making up the next two largest groups with 15.5% and 15.4% of the population respectively. Across racial and ethnic categories, the majority of groups reported anywhere from 22%–28% of the population gambled in the last 12 months with the exception of African Americans and American Indian/Alaska Natives reporting 35.7% and 40.1% respectively.

In terms of health behaviors, we recorded self-reported smoking and binge drinking status. We found that 5.4% of the study population reported being a current smoker. Among the smoking population, 39.3% reported gambling in the last 12 months and 60.7% did not. Similarly, 16.4% of the study population reported binge drinking. Among those individuals, 33.3% reported gambling in the last 12 months, and 66.8% did not. Overall, across all of the sociodemographic factors we analyzed, there were consistent distributions of those who reported gambling between groups within the same overarching category, such as across, gender, racial/ethnic groups and across health behavior groups.

Table 1 shows the socio-demographic breakdown of the study population based on age, gender, race/ethnicity, health conditions, health behaviors, and social factors. Sixteen participants’ responses were dropped from the depression/anxiety category due to them electing not to answer the associated questions.

Table 1.

The socio-demographic breakdown of the study population based on age, gender, race/ethnicity, health conditions, health behaviors, and social factors. 16 participants’ responses were dropped from the depression/anxiety category due to them electing not to answer the associated questions.

Variable	Total Sample 21,671	Gambling (Yes) 5,762 (26.6%)	Gambling (No) 15,909 (73.4%)
Demographics
- Mean Age (SD)	52.5 (17.1)	53.2 (15.9)	52.3 (17.5)
- Male (%)	9,279 (42.8)	2,675 (28.8)	6,604 (71.2)
- Female (%)	12,392 (57.2)	3,087 (24.9)	9,305 (75.1)
Race/Ethnicity
- White (%)	10,719 (49.5)	2,815 (26.3)	7,904 (73.7)
- African American (%)	1,023 (4.7)	365 (35.7)	658 (64.3)
- Latino (%)	3,333 (15.4)	904 (27.1)	2,429 (72.9)
- Asian (%)	3,364 (15.5)	759 (22.6)	2,605 (77.4)
- American Indian/Alaska native (%)	162 (0.8)	65 (40.1)	97 (59.9)
- Other single/multiple race (%)	3,070 (14.2)	854 (27.8)	2,216 (72.2)
Health Conditions
- Hypertension (%)	8,753 (40.4)	2,641 (30.2)	6,112 (69.8)
- Diabetes (%)	2,809 (13)	884 (31.5)	1,925 (68.5)
- Depression/Anxiety (%) *	2,541 (11.7)	644 (25.3)	1,897 (74.7)
Health Behaviors
- Current Smoker (%)	1,172 (5.4)	461 (39.3)	711 (60.7)
- Binge Alcohol Use (%)	3,561 (16.4)	1,184 (33.3)	2,377 (66.8)
Social Factors
- College degree or Above (%)	12,122 (55.9)	2,955 (24.4)	9,167 (75.6)
- Insurance (%)	20,991 (96.9)	5,606 (26.7)	15,385 (73.3)

^*16 participants were dropped based on “proxy skipped” response

The first logistic regression model, Table 2, which looks at hypertension suggests that gender, age, and gambling in the past 12 months are significant predictors of the outcome variable hypertension (HTN). Females are less likely to experience HTN compared to males, and older individuals are more likely to experience HTN compared to younger individuals. Additionally, individuals who have gambled in the past 12 months are 1.35 times more likely to experience HTN when compared to those who have not (OR: 1.35) 95% CI [1.24, 1.48]. However, there is no significant interaction between gender and gambling, suggesting that the effect of gambling on the outcome of HTN is similar for both males and females.

Table 2.

Depicts the outcomes of the logistic regression investigating the relationship between gambling habits, gender, and age on hypertension. The table shows the crude odds ratio (OR), as well as the OR adjusted for gender and age. Results were evaluated at a 95% confidence interval.

	Crude OR	OR adjusted for gender	OR adjusted for gender and age	OR adjusted and with interaction
Outcome (y): Hypertension	Model 1 OR [95% CI]	Model 2 OR [95% CI]	Model 3 OR [95% CI]	Model 4 OR [95% CI]
Gambling	1.36 [1.28,1.44]	1.33 [1.25,1,42]	1.35 [1.27,1.44]	1.35 [1.24,1.48]
Gender		0.66 [0.62,0.69]	0.68 [0.64,0.72]	0.68 [0.63,0.73]
Age			1.05 [1.05,1.05]	1.05 [1.05,1.05]
Gambling*Gender				1.0 [0.88,1.14]

When we include an adjustment for smoking, binge drinking, and education level, we see that gambling in the past 12 months was associated with a 34.04% increase in the odds of having hypertension (OR = 1.3404, p < 0.001). Current smoking was associated with a 15.21% reduction in the odds of hypertension (OR = 0.8479, p = 0.007), indicating a significant but inverse relationship. Binge drinking in the past 30 days was associated with a 26.50% increase in the odds of hypertension (OR = 1.2645, p < 0.001). Educational attainment showed a protective effect, as individuals with a college degree or higher had 29.06% lower odds of hypertension compared to those without a college degree (OR = 0.7094, p < 0.001).

Table 2 depicts the outcomes of the logistic regression investigating the relationship between gambling habits, gender, and age on hypertension. The table shows the crude odds ratio (OR), as well as the OR adjusted for gender and age. Results were evaluated at a 95% confidence interval.

Table 3 depicts the results of the next logistic regression model which suggests that gender, age, and gambling in the past 12 months are significant predictors of diabetes. Females are less likely to develop diabetes compared to males, and older individuals are more likely to develop diabetes compared to younger individuals. Additionally, individuals who have gambled in the past 12 months are 1.32 times more likely to develop diabetes (OR: 1.32) 95% CI [1.16, 1.49]. However, there is no significant interaction between gender and gambling, suggesting that the effect of gambling on diabetes is similar for both males and females.

Table 3.

Depicts the outcomes of the logistic regression investigating the relationship between gambling habits, gender, and age on diabetes. The table shows the crude odds ratio (OR), as well as the OR adjusted for gender and age. Results were evaluated at a 95% confidence interval.

	Crude OR	OR adjusted for gender	OR adjusted for gender and age	OR adjusted andwith interaction
Outcome (y): Diabetes	Model 1 OR [95% CI]	Model 2 OR [95% CI]	Model 3 OR [95% CI]	Model 4 OR [95% CI]
Gambling	1.32 [1.21,1.43]	1.3 [1.19,1.41]	1.31 [1.2,1.43]	1.32 [1. 16,1.49]
Gender		0.72 [0.67,0.78]	0.76 [0.7,0.83]	0.76 [0.69,0.84]
Age			1.04 [1.03,1.04]	1.04 [1.03,1.04]
Gambling*Gender				0.99 [0.83,1.18]

When controlling for smoking, binge drinking, and education level, the analysis shows that individuals who engaged in gambling in the past 12 months are 31.8% more likely to have diabetes compared to those who did not gamble, after controlling for the other variables (OR = 1.3178, p < 0.001). In contrast, current smoking status appears to be associated with lower odds of diabetes. Specifically, smokers have 14.98% lower odds of having diabetes compared to non-smokers, although this result is marginally significant (OR = 0.8502, p = 0.055), indicating a weak association. Binge drinking in the past 30 days is strongly associated with an increased likelihood of diabetes. Those who engaged in binge drinking are 129.8% more likely to have diabetes compared to non-binge drinkers (OR = 2.2978, p < 0.001). Finally, having a college degree or higher is associated with a 42.43% reduction in the odds of having diabetes, suggesting that higher educational attainment has a protective effect (OR = 0.5757, p < 0.001).

Table 3 depicts the outcomes of the logistic regression investigating the relationship between gambling habits, gender, and age on diabetes. The table shows the crude odds ratio (OR), as well as the OR adjusted for gender and age. Results were evaluated at a 95% confidence interval.

Discussion

The results of this study are consistent with previous research connecting gambling behaviors to chronic illnesses. Specifically, gambling was significantly associated with increased odds of hypertension and diabetes. Thus, our findings aligned with studies that highlight the physiological impacts of gambling, such as elevated cortisol levels and heart rate.¹⁰ Overall, our findings indicate a correlation between gambling behaviors and chronic physical illnesses.

In addition, several limitations of this study must be acknowledged. First, the CHIS dataset relied on self-reported data (either online or over the phone), which introduces the potential for recall bias, confusion, or inaccuracies in responses. Second, the cross-sectional design prevents the ability to establish causal relationships between gambling and health outcomes. Third, the findings may have limited generalizability to the broader population, as the data was solely collected from California residents. Fourth, a small percentage of participants did not respond to certain questions in the survey, and we chose to exclude these individuals from the analysis. This decision may have introduced bias into our results, as the reasons for their non-responses remain unknown. Last, we only considered a limited range of psychosocial variables in this study, focusing primarily on gender and age, leaving other potential influencers such as socioeconomic status unexamined. Nevertheless, despite these limitations, our findings contribute valuable insights into the health implications of gambling.

Future research could address these gaps by employing longitudinal designs to establish causal relationships between gambling and health outcomes. Additionally, investigating specific gambling behaviors, such as frequency, type, or expenditure, could provide a more nuanced understanding of how gambling affects physical and mental health. Expanding the scope of psychosocial factors may help us to further understand the complex relationship of factors that influence health outcomes. Lastly, it is imperative to compare results across different populations and regulatory environments in order to ensure the generalizability of results and better understand the role of external factors in gambling-related health risks.

The findings of this study have important implications for future public health interventions and policies. For example, targeted education campaigns could distribute educational materials and resources about chronic diseases throughout the communities identified as at risk. These campaigns could include educating pathological gamblers on the signs and symptoms of chronic illnesses, such as diabetes and hypertension, as well as the risks of not treating them properly. Educating individuals who gamble or have a gambling problem about their increased risks for disease opens the door for them to take preventative measures. Additionally, the implementation of stress management programs tailored to individuals who gamble may help to reduce some of the physiological impacts of gambling-related stress. Policymakers should also consider stricter regulations on gambling advertisements, particularly those targeting vulnerable populations, in order to reduce exposure and potential harm. Lastly, incorporating more questions about gambling behaviors into broader health surveys could also facilitate increased data collection and better inform public health initiatives aimed at addressing the negative health impacts of gambling.