| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Psychiatry (B.J.M., R.H.P., N.M.P.), University of Connecticut Health Center, Farmington, Connecticut; the Department of Psychiatry (C.B., D.H.), Columbia University, New York, NY; and the Laboratory of Epidemiology and Biometry, Division of Intramural Clinical and Biological Research (B.F.G.), National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, Maryland.
Address correspondence and reprint requests to Nancy M. Petry, PhD, Department of Psychiatry, University of Connecticut Health Center, 263 Farmington Avenue, Farmington, CT 06030-3944. E-mail: petry{at}psychiatry.uchc.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Method: A total of 43,093 adults aged 18 years and older were evaluated in the 2001 to 2002 National Epidemiologic Survey on Alcohol and Related Conditions. Self-reported medical diagnoses and past-year medical services used were assessed.
Results: Pathologic gamblers were more likely than low-risk individuals to have been diagnosed with tachycardia (odds ratio [OR] = 1.77; 95% confidence interval [CI] = 1.05–2.97), angina (OR = 2.35; 95% CI = 1.33–4.15), cirrhosis (OR = 3.90; 95% CI = 1.11–13.72), and other liver disease (OR = 2.98; 95% CI = 1.07–8.26). Gambling severity was also associated with higher rates of medical utilization with pathologic gamblers more likely than low-risk individuals to have been treated in the emergency room in the year before the survey (OR = 1.98; 95% CI = 1.27–3.09). Significant effects of gambling severity remained even after controlling for demographic characteristics (age, gender, ethnicity, marital status, education, income, and region of the country) and behavioral risk factors such as body mass index, alcohol abuse and dependence, nicotine dependence, and mood and anxiety disorders.
Conclusions: A lifetime diagnosis of pathologic gambling is associated with several medical disorders and increased medical utilization, perhaps leading to a burden on healthcare costs in the United States.
Key Words: pathologic gambling • health functioning • medical utilization • epidemiology • psychiatric disorder • mental health
Abbreviations: AUDADIS-IV = Alcohol Use Disorder and Associated Disabilities Interview Schedule-IV; BMI = body mass index; CI = confidence interval; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders–Fourth Edition; ER = emergency room; NESARC = National Epidemiologic Survey on Alcohol and Related Conditions; NIAAA = National Institute on Alcohol Abuse and Alcoholism; OR = odds ratio; SES = socioeconomic status; SF-12v2 = Short-Form Health Survey, Version 2.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Pathologic gambling is related to a range of psychiatric comorbidities, including anxiety and depressive disorders (4,8), alcohol and substance use (3,6), and cigarette smoking (8,9). In addition, evidence is emerging that pathologic gambling may be associated with adverse health consequences. According to the DSM-IV (1), individuals diagnosed with pathologic gambling "may be prone to developing general medical conditions that are associated with stress (e.g., hypertension, peptic ulcer disease, migraine)" (p. 616). However, no published study has evaluated medical conditions among pathologic gamblers.
A prior report (10) found that pathologic gamblers had high rates of "severe physical symptoms" (p. 280) such as fatigue, colds and influenza, headaches, gastric pain, and nausea. However, this was a purely descriptive study with no comparison group.
Some studies have examined the relationship between gambling severity and self-evaluation of quality of life. A study with primary care patients found that increased severity of lifetime gambling problems was associated with poorer ratings of general health (11), but health functioning was limited by assessment with a single unstructured item with unknown psychometric properties. Subsequent research conducted with community volunteers (12,13), senior citizens (14,15), and in primary care medical (16) and dental (17) clinics extended these findings using brief standardized measures of quality of life (e.g., Medical Outcomes Study Short Form-36 (18)) and psychopathology (e.g., Beck Depression Inventory (19)). Individuals with more severe lifetime gambling problems consistently rated aspects of their current physical (e.g., impact of bodily pain on daily functioning, activities are limited as a result of health) and mental (e.g., depression) health more poorly than individuals without lifetime gambling-related problems. A limitation of these studies, however, was that self-appraisal assessed health status. Appraisal of health is not necessarily congruent with actual functioning or onset of medical conditions (20). To better understand the relationship between gambling behaviors and health, the prevalence of specific medical conditions should be compared among individuals with varying degrees of gambling severity.
Because pathologic gambling is associated with a number of stressors, including financial losses, fractured marital relations, and employment difficulties (7), disorders with a stress-related component such as hypertension may be increased in problem and pathologic gamblers. A higher rate of cardiac disease, including sudden cardiac death, may be present among groups of gamblers with stress or stress responsivity representing possible mediating factors (21,22). Heart disease is also associated with history of cigarette smoking, which is frequently comorbid with pathologic gambling (8,9). Medical disorders linked with alcohol and drug abuse such as cirrhosis may also occur at high rates, because pathologic gamblers have elevated rates of substance use disorders (3,6). Furthermore, the sedentary nature of gambling may be especially appealing to individuals with physically limiting disorders such as arthritis or obesity. Greater rates of these medical conditions may be subsequent to common risk factors for gambling and medical disorders or they may be uniquely related to pathologic gambling.
Given the rapid proliferation of legalized gambling, and the hypothesized association between pathologic gambling and health functioning (23), a better understanding of this relationship is needed. To expand on prior research, we provide nationally representative data on the prevalence and co-occurrence of gambling disorders and self-reported medical diagnoses evaluated in the 2001 to 2002 National Institute on Alcohol Abuse and Alcoholism (NIAAA) National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). This is the largest prevalence study of psychiatric disorders ever conducted (24,25) and provides a rich opportunity to identify current health correlates associated with lifetime gambling behavior. Additionally, if lifetime problem and pathologic gambling are associated with any health conditions, they may also be related to increased use of medical resources such as emergency room visits and hospitalizations.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Data were weighted to reflect design characteristics of the NESARC and to account for oversampling. Adjustment for nonresponse across variables was performed at the household and person level. Weighted data were adjusted to be representative of the civilian population of the United States on a variety of sociodemographic variables, including region of the country, age, race–ethnicity, and sex based on the 2000 Decennial Census.
Assessment of Gambling
Gambling was defined in the NESARC as playing cards for money, betting on horses or dogs or sports games, playing the stock or commodities market, buying lottery tickets, or playing bingo or KENO or gambling at a casino, including playing slot machines. DSM-IV criteria (1) were used to classify respondents with pathologic gambling as assessed by the NIAAA Alcohol Use Disorder and Associated Disabilities Interview Schedule-IV (AUDADIS-IV), a structured diagnostic interview designed for use by lay interviewers (26). Gambling behavior was evaluated with 15 structured questions, which operationalized the 10 DSM-IV criteria. Internal consistency of the symptom items and criteria for pathologic gambling were excellent (4) (
’s = 0.92 and 0.80, respectively). Test–retest reliability of lifetime symptoms was also excellent (27) (intraclass correlation coefficient = 0.76, 95% confidence interval [CI] = 0.68–0.82).
Appraisal of Physical and Mental Health Functioning
The Short-Form Health Survey, version 2 (SF-12v2 (28)) assessed physical and emotional functioning in the past 4 weeks. Two component summary scores (Physical Health and Mental Health) and eight index scores (physical functioning, role–physical, bodily pain, general health, vitality, social functioning, role–emotional, and mental health) are produced. Psychometric analysis of the SF-12v2 revealed a 2-week test–retest reliability of 0.76 and convergent validity of 0.67 to 0.97 (29). Higher scores on the measure indicate better functioning. The measure has mean summary and index scores of 50 with standard deviations of 10 (28,29).
Assessment of Medical Diagnoses
Respondents were asked whether they experienced a list of specific medical conditions in the past year. Diagnoses assessed included hypertension, tachycardia, arteriosclerosis, angina, myocardial infarction, other heart disease, cirrhosis, other liver disease, stomach ulcer, gastritis, and arthritis. If respondents endorsed one of these conditions, they were asked if a physician or other health professional made the diagnosis; only diagnoses reported having been made by a physician or other health professional were considered positive. In this study, we assessed the relationship between all assessed medical conditions and gambling severity.
Medical Utilization
Respondents were asked to indicate the frequency of the following events in the past year: times stayed overnight in the hospital, days stayed in the hospital, times treated in a hospital emergency room (ER), and injuries that required medical attention or caused decrease of usual activities for more than half a day.
Assessment of Behavioral Risk Factors
Behavioral factors that may be associated with health were controlled for in analyses evaluating the relationships between gambling severity and health functioning. Risk factors included in this report were body mass index (BMI), lifetime history of alcohol abuse or dependence, lifetime history of nicotine dependence, and lifetime diagnosis of any DSM-IV mood or anxiety disorder.
BMI is a measure of body weight that accounts for height and applies to both adult women and men. BMI was computed by dividing respondents’ self-reported weight in kilograms by the square of reported height in meters. Although appraisal of height and weight is not precise, this method of computing BMI correlates highly (r >0.90) with operationally measured BMI (30–32). BMI greater than 30 kg/m2 is considered obese (33).
The AUDADIS-IV was used to diagnose lifetime alcohol abuse or dependence, nicotine dependence, mood disorders, and anxiety disorders. Test–retest reliabilities of AUDADIS-IV diagnoses were good to excellent (34,35). The validity of nicotine dependence, alcohol use disorders, and mood and anxiety disorders were also good to excellent (36–38).
Statistical Analysis
Respondents were classified into one of four gambling groups based on their responses to the AUDADIS-IV. All respondents were asked if they had ever gambled five or more times in a single year. This item served as a "gatekeeper" for the assessment of gambling behavior, because individuals who responded "no" were not administered the 15 structured gambling-related questions. Respondents who reported never gambling five or more times in 1 year were termed "low-risk individuals" (this group includes both never gamblers and individuals who may have gambled during their lifetimes, although never more than five times per year; unfortunately, the AUDADIS-IV did not assess whether participants had ever wagered so this group could not be further subdivided into never gamblers). Respondents endorsing five or more DSM-IV criteria were classified as "pathologic gamblers" and those that endorsed three or four DSM-IV symptoms were termed "problem gamblers." Respondents who endorsed gambling five or more times in any single year, but met two or fewer DSM-IV symptoms, were termed "at-risk gamblers." These gambling classifications are consistent with prior research (3).
2 tests evaluated differences on demographic variables across the four groups. Linear regression analyses assessed differences in SF-12v2 scores by gambling status with SF-12v2 scores entered as dependent variables, and gambling status, demographic variables (age, gender, race–ethnicity, marital status, income, education, and region of the country), and behavioral risk factors (BMI, lifetime alcohol abuse or dependence, lifetime nicotine dependence, and lifetime mood or anxiety disorders) included as independent variables. Age and BMI were entered as continuous variables and the remaining were categorical.
Odds ratios (ORs), derived from a series of logistic regression analyses, evaluated associations between gambling severity and past-year medical diagnoses. Three models were estimated: model 1 did not control for the effects of any confounding variables, model 2 controlled for demographic variables (age, gender, race–ethnicity, marital status, income, education, and region of the country), and model 3 controlled for demographic variables and behavioral risk factors (BMI, lifetime alcohol abuse or dependence, nicotine dependence, and mood or anxiety disorders). We also considered controlling for substance use disorders (SUDs) and antisocial personality disorder; however, given the high overlap between these disorders and those already in the analyses,* inclusion of these variables would be inappropriate because of multicollinearity and would inflate confidence intervals of parameter estimates.
ORs also evaluated associations between gambling severity and medical utilization. Given the nonnormal distribution of these data, responses were recoded to dichotomous measures to indicate whether any of the medical events occurred in the last 12 months. Similar to the analyses for medical diagnoses, three separate models were estimated. Data were analyzed using SUDAAN (39), a software package that uses Taylor series linearization to adjust for the design effects of the complex sampling methodology of this survey.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
The prevalence of lifetime pathologic gambling in this sample was 0.42%. In addition, 0.90% of respondents were classified as problem gamblers (meeting three or four DSM-IV criteria). At-risk gamblers and low-risk individuals comprised 25.84% and 72.84% of the full sample, respectively. Table 1 displays demographic characteristics of respondents in each of the four groups. Statistically significant differences were noted on all demographic variables. Table 1 also displays information about the relationship between gambling severity and behavioral risk factors. Increased gambling severity was associated with current obesity status, alcohol abuse or dependence, nicotine dependence, and mood and anxiety disorders.
|
Linear regression analyses evaluated whether gambling severity was associated with SF-12v2 scores (Table 2). Because each of the demographic characteristics and behavioral risk factors differentiated between the gambling groups, they were included in the analyses as covariates. After controlling for covariates, gambling classification was associated with both the Physical Health and Mental Health Component Summary scores (p < .05). Analyses with the eight index scores revealed that gambling severity was associated with each aspect of physical and mental health except the general health and vitality scales. For the significant indices, increased gambling severity was associated with decreased appraisal of health status. Table 2 also reports standard effect sizes comparing self-reported health functioning of pathologic gamblers versus low-risk individuals. The effect sizes indicate that, although statistically significant, the magnitude of effects is small.
|
The ORs of gambling severity and medical diagnoses are shown in Table 3. In model 3, which controlled for demographic characteristics and behavioral risk factors, pathologic gamblers were significantly more likely than low-risk individuals to have been diagnosed with four of the 11 conditions assessed: tachycardia, angina, cirrhosis, and other liver disease. In addition, at-risk gamblers and problem gamblers were more likely than low-risk individuals to have been diagnosed with hypertension. Problem gamblers were more likely than low-risk individuals to also have been diagnosed with angina and cirrhosis. The only diagnosis that conferred less risk was cirrhosis; at-risk gamblers were less likely than low-risk individuals to be diagnosed with cirrhosis, but as noted previously, problem and pathologic gamblers had increased risk of this disorder.
|
Table 4 displays ORs of gambling severity and medical utilization. In model 3, which controlled for demographic characteristics and behavioral risk factors, pathologic gamblers were more likely than low-risk individuals to have been treated in the ER. Relative to low-risk individuals, problem gamblers and at-risk gamblers reported an increased likelihood of having been treated in the ER and experiencing a severe injury.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
The causal relationship between gambling severity and certain medical conditions is poorly understood. Other factors may affect the relationship between pathologic gambling and some medical conditions. Low socioeconomic status (SES) and obesity are associated with increased rates of a range of health problems (40,41), including cardiovascular disease (42–44) and arthritis (45). Alcohol and other substance use disorders may also lead to certain medical disorders, particularly cirrhosis and other liver disease. Pathologic gambling is also highly comorbid with substance disorders (3,4,6). However, the relationship between pathologic gambling and these medical disorders remained even after controlling for income, BMI, nicotine dependence, alcohol use disorders, and mood and anxiety disorders.
Other factors likely mediate the relationships between pathologic gambling and some medical conditions. The sustained stress of casino gambling may lead to cardiac disease (46). Onset of medical symptoms may also be the result of, at least in part, behavioral factors (47). Individuals with a gambling disorder may be more sedentary, experience more stress (46), breath more secondhand smoke (8,9), and participate in fewer health-related activities, which could serve as risk factors for developing serious and chronic medical conditions. At the same time, individuals with some medical conditions and physical limitations may also be more likely to gamble, because gambling is a form of recreation that requires minimal physical activity. Finally, other etiologies that were not assessed in the NESARC such as genetic factors (13,48,49), may exist for some of these medical disorders, and they may be independently or additively associated with problem and pathologic gambling.
The increased rates of medical problems identified in this study among problem and pathologic gamblers were paralleled in the higher rates of medical utilization seen in these groups. At-risk, problem, and pathologic gamblers were more likely than low-risk individuals to have been treated in an ER, and at-risk and problem gamblers were more likely to have experienced a severe injury that impaired daily functioning. This is the first study to report on medical utilization among problem and pathologic gamblers, and it suggests that gambling problems may affect healthcare costs and medical expenditures.
Some prior research has suggested that recreational gambling may be associated with health benefits in older adults (50). However, findings from the present study indicate that gambling even five times a year (i.e., at-risk gambling) can be associated with adverse health consequences. Persons classified as at-risk gamblers in this survey were significantly more likely than low-risk individuals to have been diagnosed with hypertension, and they were more likely to have received treatment in the ER and have experienced a severe injury in the past year. At-risk gamblers were also more likely than their low-risk counterparts to be obese and have a history of a mood or anxiety disorder, an alcohol use diagnosis, and nicotine dependence. Although risk of cirrhosis was lowest in the at-risk gamblers, the overall rate of cirrhosis was less than 1% in the entire sample. Taken together, these findings indicate that even a moderate amount of gambling (five or more times in a year) is associated with some decreased health functioning and increased medical utilization when the sample is not limited to older adults. Because the at-risk gambling group comprises approximately one fourth of the population, public health implications of gambling may exist.
In total, these data indicate that lifetime gambling problems are associated with current health functioning and medical utilization. Strengths of this study include the large sample size that was randomly selected from the U.S. population and high response rate. The interviews used a structured diagnostic instrument, were conducted face-to-face, and psychiatric diagnoses were made according to DSM-IV criteria. Medical conditions reported as diagnosed by a medical professional, rather than appraisal of health functioning, were used as outcome measures. In addition, numerous possible confounding variables were controlled in the analyses, providing a clearer picture of the relationship between gambling severity and health outcomes.
Limitations to this study also exist. First, the reference group for the odds ratios included nongamblers and low-frequency gamblers. Because the AUDADIS-IV did not assess whether participants had ever gambled, they were combined in a single group (low-risk individuals). Other limitations relate to the definitions used for at-risk and problem gambling. Respondents who reported gambling five or more times in a single year and had zero, one, or two DSM-IV symptoms were classified as at risk and those with three or four symptoms as problem gamblers. Results may have differed somewhat with other cutoff points. Although debate exists regarding gambling classification (51), the method used in this study was consistent with prior literature (3,52,53) and provides a conservative test of the relationship between gambling severity and health functioning.
Another limitation is that different timeframes were used for the assessment of gambling severity, medical diagnoses, and medical utilization. Lifetime gambling behavior was analyzed, whereas data on medical diagnoses and health service utilization were for the past year only. The use of lifetime gambling behavior may, however, be a strength, because the results indicate that even prior gambling behavior can confer later health problems.
Finally, not all aspects of health functioning were evaluated, and physicians or physical tests did not independently confirm the diagnoses reported in this study. All participants, however, indicated that a physician or other health professional initially made the diagnoses analyzed as part of this study. Additionally, we relied on self-report of medical utilization, which may not be reliable in certain subgroups such as individuals with alcohol use disorders (54).
In summary, this study confirms prior research indicating lower appraisal of physical and mental health functioning among pathological gamblers (11–17), and it is the first to document a risk of specific medical diagnoses and increased medical service utilization among individuals with gambling problems. Problem and pathologic gamblers were significantly more likely than low-risk individuals to rate their health more poorly and to be diagnosed with certain medical conditions. Problem and pathologic gamblers were also more likely to have received expensive medical services in the prior year. These results suggest additional adverse personal and financial hardships associated with gambling, and these findings warrant more research and clinical attention directed toward disordered gambling in medical settings.
We thank Dr. Richard Feinn for assistance with data management and the U.S. Census Bureau field representatives who administered the NESARC interview.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Dr. Morasco is now affiliated with the Portland VA Medical Center.
Received for publication January 6, 2006; revision received June 16, 2006.
DOI:10.1097/01.psy.0000238466.76172.cd
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
