- Split View
-
Views
-
Cite
Cite
Rosa Alati, Stuart Kinner, Jake M. Najman, Greg Fowler, Kerrianne Watt, David Green, GENDER DIFFERENCES IN THE RELATIONSHIPS BETWEEN ALCOHOL, TOBACCO AND MENTAL HEALTH IN PATIENTS ATTENDING AN EMERGENCY DEPARTMENT, Alcohol and Alcoholism, Volume 39, Issue 5, September/October 2004, Pages 463–469, https://doi.org/10.1093/alcalc/agh080
- Share Icon Share
Abstract
Aims: There is evidence of a non-linear relationship between alcohol consumption and mental health status, and of an association between tobacco use and poor mental health. This paper examines the nature of the association between usual alcohol consumption, tobacco use and symptoms of anxiety and depression in Emergency Department patients in Queensland, Australia. Methods: A cross sectional survey of patients aged 16–84 presenting for treatment over a 14 day period to Gold Coast Hospital Emergency Department using socio-demographic items, the Alcohol Use Disorders Identification Test (AUDIT) to measure moderate, hazardous and harmful alcohol consumption, and the Hospital Anxiety and Depression Scale (HADS) to measure state anxiety and depression. Results: 812 patients were interviewed. Gender differences in results were evident. For men, there was a U-shaped relationship between alcohol consumption and anxiety/depression, and a linear association between smoking and anxiety. For women, alcohol consumption and anxiety/depression showed a more linear relationship, but there was no significant relationship between tobacco use and anxiety/depression. Conclusions: There may be important gender differences in the relationships between alcohol consumption, tobacco use and mental health status. This study supports previous evidence that mental health status of non-drinkers is worse than that of moderate drinkers, but only among males.
(Received 7 February 2004; first review notified 12 April 2004; in revised form 21 May 2004; accepted 23 May 2004)
INTRODUCTION
Tobacco use and excessive alcohol consumption remain two major public health problems (Ezzati et al., 2002; World Health Organization, 2002). The evidence also suggests an association between poor mental health and substance use disorders (Power et al., 1998; Rodgers et al., 2000b; Degenhardt et al., 2000, 2001; Degenhardt and Hall, 2001). Individuals who report more symptoms of anxiety and depression are also highly represented amongst those who smoke tobacco and consume alcohol at harmful levels (Degenhardt and Hall, 2001, 2003; Caldwell et al., 2002; World Health Organization, 2002). This paper aims to further investigate the nature of the relationship between mental health and substance use problems.
Alcohol use, anxiety and depression
Harmful alcohol use and alcohol dependence are the most frequently experienced substance use disorders in Australia (Degenhardt and Hall, 2003). Amongst those diagnosed with substance use disorders, alcohol disorders are three times more common than illicit drug disorders (Henderson et al., 2000) and their co-occurrence with anxiety and affective disorders is high (Degenhardt and Hall, 2001, 2003). While the association between harmful alcohol use and mental health problems is well established, the precise nature of the relationship between alcohol use in general and mental health remains unclear. Some recent studies have suggested a non-linear or ‘U-shaped’ relationship between anxiety and depression and the use of alcohol (Power et al., 1998; Rodgers et al., 2000b; Caldwell et al., 2002), such that both hazardous/harmful drinking and abstinence from alcohol are associated with higher rates of depression and anxiety than moderate/non-hazardous drinking (Rodgers et al., 2000b; Degenhardt and Hall, 2001; Caldwell et al., 2002).
These findings are reminiscent of the literature on alcohol consumption and physical health where moderate drinkers, particularly males, live longer and appear to have a variety of health advantages, when compared with non-drinkers (World Health Organization, 2002). Whereas these health benefits are often attributed to the protective effect of alcohol on cardiovascular activity (National Health and Medical Research Council, 2001), the reasons for a non-linear association between alcohol consumption and anxiety and depression are less clear. One suggestion is that alcohol may have a protective effect on mental health—that the lower symptoms of anxiety and depression amongst moderate drinkers may reflect the better societal adjustment that moderate drinkers experience as a result of their patterns of alcohol consumption (Rodgers et al., 2000a). Another view is that among alcohol abstainers there may exist a distinct subgroup who have had an alcohol problem in the past, and who are characterized by the same mental health patterns as those who drink heavily (Goldman and Najman, 1984; Rodgers et al., 2000b). A second ‘artefactual’ explanation for the observed non-linear relationship between alcohol consumption and mental health is that non-drinkers may share with heavy drinkers a variety of circumstances, such as low socio-economic status (Rodgers et al., 2000a), which are known to correlate positively with poorer health.
Tobacco use, anxiety and depression
High levels of alcohol use are often associated with tobacco use. In Australia, those who drink are between 1.8 and 4.2 times more likely to smoke tobacco, with those meeting the DSM-IV criteria for alcohol disorder being much more likely to smoke tobacco (Degenhardt and Hall, 2003). Smoking is therefore positively associated with heavy alcohol use. Given that alcohol disorders are associated with anxiety and depression, a similar association may exist between tobacco use and mental health.
Although evidence for this association is limited at present, two population surveys have supported a strong association between regular tobacco use and poorer mental health in the general population. In a stratified probability sample of 10 641 adults, representative of the Australian population, Degenhardt and Hall (2001) found that regular tobacco users were significantly more likely than non-smokers to experience both alcohol disorders and higher rates of anxiety and affective disorders. In another large (n = 2725) community sample of Australian adults, Jorm et al. (1999) found a significant association between tobacco use and both anxiety and depression. In both studies, these relationships did not appear to be a consequence of confounding.
If smoking and mental health are indeed associated, one question of interest is whether the nature of this association is similar to that observed between alcohol and mental health. If this were the case, then one might expect moderate tobacco consumption to confer some mental health benefits, with both non-smokers and heavy smokers characterized by relatively poorer mental health. However this would counter findings from the mortality literature, in which no health benefits are gained in consuming tobacco, and a negative linear relationship is found between tobacco use and physical health status (World Health Organization, 2002). It is therefore of some interest to examine differences in mental health status across the whole spectrum of tobacco consumption.
Emergency Departments are a potential first point of contact for a range of health problems, and are thus an excellent opportunity for screening and early identification. Hazardous and harmful drinking are common among patients of Emergency Departments both in Australia and overseas (Spirito et al., 2001; Cassar et al., 2002; Tait et al., 2002; Woolfenden et al., 2002; Barnett et al., 2003; Rockett et al., 2003). The interplay between alcohol or tobacco use and mental health has received relatively little attention in an Australian Emergency Department context. Evidence of a nonlinear relationship between alcohol or tobacco use and mental health, in Emergency Department populations, would carry important implications for public health research and intervention.
The following paper has three aims. First, it will examine the prevalence of alcohol consumption, tobacco use, and symptoms of anxiety and depression in a sample of patients presenting to an Emergency Department in a large, metropolitan hospital in south-east Queensland, Australia. Second, it will explore the nature of the relationship between anxiety/depression and alcohol consumption in an Emergency Department setting. Third, it will examine the nature of any association between anxiety/depression and tobacco use in this population.
METHODS
Study design and setting
A cross-sectional survey was conducted 24 h per day over a 14-day period in October 2002. Participants were patients who presented for treatment at the Gold Coast Hospital Emergency Department (ED). The Gold Coast is a densely populated area and a major destination for both Australians and overseas tourists. The Gold Coast Hospital is the only public hospital in the region and is situated opposite a private hospital. It is therefore more likely to attract patients who are less wealthy and may not have access to private health insurance.
Sample
A sequential list of all eligible patients (in order of presentation to the hospital) was generated from hospital computer records. Eligible patients were aged 16–84 years, able to be interviewed and able to give informed consent. Patients were approached for interview on the basis that they met these eligibility criteria. Previous research carried out in this setting (Roche et al., 2001) suggested that the resources available would not enable the interviewers to approach every patient for interview. Over the course of the study, four out of every five eligible patients were successfully interviewed. In situations where more than one eligible patient was available for interview at a given time, patients were approached in the order in which they had presented, with the non-selected patient potentially lost to interview. Consent was obtained from patients for participation in a 20 min, face-to-face interview.
Of 1446 presenting patients, 280 were not approached, in accordance with study protocol. A further 178 patients were not eligible for participation (63 could not be interviewed due to the severity of their condition, and 115 did not wait for medical treatment). Thus, of 988 eligible patients, 812 (82.2%) were interviewed, 107 (10.8%) refused to participate, and 69 (7.0%) patients received treatment and left hospital before they could be interviewed.
Measures
Measures included participants' socio-demographic characteristics, usual level of tobacco consumption, and two standardized instruments to assess (i) levels of alcohol consumption and (ii) prevalence of anxiety and depression. Socio-demographic information included data regarding age, gender, Australian residency, indigenous status, marital status, education, household income and private health insurance.
Participants were asked to report on their current tobacco use, measured as the number of cigarettes they smoke on an average day (‘How many cigarettes do you smoke on an average day?’). For the purposes of analysis and interpretation, this continuous measure of current tobacco use was categorized into four levels of tobacco consumption, based on the reported average number of cigarettes smoked per day: non-smoker (0 cigarettes), light smoker (1–9 cigarettes), moderate smoker (10–19 cigarettes) and heavy smoker (20+ cigarettes).
Usual alcohol consumption was assessed using the Alcohol Use Disorders Identification Test (AUDIT, Babor et al., 1992). The AUDIT is a 10-item screening instrument designed to measure three domains of alcohol use: consumption (items 1–3), dependence (items 4–6) and related problems (items 7–10). AUDIT scores fall into four categories: non-drinker (score = 0), non-hazardous drinking (females: 1–6; males: 1–7), hazardous drinking (females: 7–12; males: 8–14) and harmful drinking (females: 13–40; males: 15–40). Cut-offs used in this study follow those identified by the scale authors, which were found to have a sensitivity of 95–100% in a population of 913 drinking patients (Babor et al., 1992). In the present study, the AUDIT was found to have a reliability (Cronbach's alpha) of 0.84. Blood alcohol concentration (BAC) was not measured in this study.
Mental health was assessed using the Hospital Anxiety and Depression Scale (HADS), a self-report measure designed to detect symptoms of state anxiety and depression in non-psychiatric hospital settings (Zigmond and Snaith, 1983). The scale contains seven depression items (four reverse scored) and seven anxiety items, rated on a four-point scale. On each subscale, a score of 7 or less is considered a non-case, 8–10 a doubtful case, and 11 or more a definite case (Zigmond and Snaith, 1983). In the present sample the scales had a reliability (Cronbach's alpha) of 0.83 (anxiety) and 0.78 (depression).
Covariates used in the multivariate analysis included age, education, income and urgency of case at presentation (measured by using the Australian Triage Scale) (ACEM, 2004). Age ranged from 16 to 84, education from primary school to university, income from less than $10 000 per annum to more than $50 000 per annum, and Triage Code from more urgent (2) to less urgent (5). All covariates were treated as continuous variables. The Triage Code allocated to patients on presentation by hospital staff was obtained from the Emergency Department patient records system.
RESULTS
Over half (n = 452; 56%) of the sample was male and the average age was 40.11 years (range 16–84 years). Over 95% of participants were Australian residents and fewer than 3% identified as indigenous; roughly equal numbers were single (41%) and married or de facto (42%). At the time of presentation just over half (54%) were employed and fewer than one in five (15.3%) reported a total household income in excess of $50 000. Only 13% of participants reported holding private health insurance (see Table 1).
. | Males (n = 452) . | Females (n = 360) . | Full sample (n = 812) . | |||
---|---|---|---|---|---|---|
Age (mean years) | 40.88 | 39.15 | 40.11 | |||
Australian resident (%) | 95 | 94 | 95 | |||
Indigenous (%) | 3 | 3 | 3 | |||
Marital status (%) | ||||||
Single | 45 | 36 | 41 | |||
Married/de facto | 42 | 43 | 42 | |||
Separated/divorced/widowed | 13 | 21 | 17 | |||
Main source of income (%) | ||||||
Full-time work | 45 | 26 | 37 | |||
Part-time/casual work | 14 | 22 | 17 | |||
Dole or pension | 29 | 25 | 27 | |||
Home duties | 1 | 13 | 6 | |||
Other | 12 | 14 | 13 | |||
Highest level of education completed (%) | ||||||
Primary | 9 | 9 | 9 | |||
Year 10 | 34 | 40 | 37 | |||
Year 12 | 22 | 25 | 24 | |||
TAFE or trade | 22 | 16 | 19 | |||
University | 12 | 10 | 11 | |||
Total annual household income (%) | ||||||
$0 – $20,000 | 38 | 49 | 43 | |||
$20 000 – $40 000 | 32 | 32 | 32 | |||
$40 000 or more | 30 | 19 | 25 | |||
Private health insurance (%) | 14 | 12 | 13 |
. | Males (n = 452) . | Females (n = 360) . | Full sample (n = 812) . | |||
---|---|---|---|---|---|---|
Age (mean years) | 40.88 | 39.15 | 40.11 | |||
Australian resident (%) | 95 | 94 | 95 | |||
Indigenous (%) | 3 | 3 | 3 | |||
Marital status (%) | ||||||
Single | 45 | 36 | 41 | |||
Married/de facto | 42 | 43 | 42 | |||
Separated/divorced/widowed | 13 | 21 | 17 | |||
Main source of income (%) | ||||||
Full-time work | 45 | 26 | 37 | |||
Part-time/casual work | 14 | 22 | 17 | |||
Dole or pension | 29 | 25 | 27 | |||
Home duties | 1 | 13 | 6 | |||
Other | 12 | 14 | 13 | |||
Highest level of education completed (%) | ||||||
Primary | 9 | 9 | 9 | |||
Year 10 | 34 | 40 | 37 | |||
Year 12 | 22 | 25 | 24 | |||
TAFE or trade | 22 | 16 | 19 | |||
University | 12 | 10 | 11 | |||
Total annual household income (%) | ||||||
$0 – $20,000 | 38 | 49 | 43 | |||
$20 000 – $40 000 | 32 | 32 | 32 | |||
$40 000 or more | 30 | 19 | 25 | |||
Private health insurance (%) | 14 | 12 | 13 |
. | Males (n = 452) . | Females (n = 360) . | Full sample (n = 812) . | |||
---|---|---|---|---|---|---|
Age (mean years) | 40.88 | 39.15 | 40.11 | |||
Australian resident (%) | 95 | 94 | 95 | |||
Indigenous (%) | 3 | 3 | 3 | |||
Marital status (%) | ||||||
Single | 45 | 36 | 41 | |||
Married/de facto | 42 | 43 | 42 | |||
Separated/divorced/widowed | 13 | 21 | 17 | |||
Main source of income (%) | ||||||
Full-time work | 45 | 26 | 37 | |||
Part-time/casual work | 14 | 22 | 17 | |||
Dole or pension | 29 | 25 | 27 | |||
Home duties | 1 | 13 | 6 | |||
Other | 12 | 14 | 13 | |||
Highest level of education completed (%) | ||||||
Primary | 9 | 9 | 9 | |||
Year 10 | 34 | 40 | 37 | |||
Year 12 | 22 | 25 | 24 | |||
TAFE or trade | 22 | 16 | 19 | |||
University | 12 | 10 | 11 | |||
Total annual household income (%) | ||||||
$0 – $20,000 | 38 | 49 | 43 | |||
$20 000 – $40 000 | 32 | 32 | 32 | |||
$40 000 or more | 30 | 19 | 25 | |||
Private health insurance (%) | 14 | 12 | 13 |
. | Males (n = 452) . | Females (n = 360) . | Full sample (n = 812) . | |||
---|---|---|---|---|---|---|
Age (mean years) | 40.88 | 39.15 | 40.11 | |||
Australian resident (%) | 95 | 94 | 95 | |||
Indigenous (%) | 3 | 3 | 3 | |||
Marital status (%) | ||||||
Single | 45 | 36 | 41 | |||
Married/de facto | 42 | 43 | 42 | |||
Separated/divorced/widowed | 13 | 21 | 17 | |||
Main source of income (%) | ||||||
Full-time work | 45 | 26 | 37 | |||
Part-time/casual work | 14 | 22 | 17 | |||
Dole or pension | 29 | 25 | 27 | |||
Home duties | 1 | 13 | 6 | |||
Other | 12 | 14 | 13 | |||
Highest level of education completed (%) | ||||||
Primary | 9 | 9 | 9 | |||
Year 10 | 34 | 40 | 37 | |||
Year 12 | 22 | 25 | 24 | |||
TAFE or trade | 22 | 16 | 19 | |||
University | 12 | 10 | 11 | |||
Total annual household income (%) | ||||||
$0 – $20,000 | 38 | 49 | 43 | |||
$20 000 – $40 000 | 32 | 32 | 32 | |||
$40 000 or more | 30 | 19 | 25 | |||
Private health insurance (%) | 14 | 12 | 13 |
AUDIT scores, HADS scores and level of tobacco consumption are presented in Table 2, as a function of gender. Slightly more than 40% of the sample used tobacco. Almost one third (30.5%) drank at hazardous/harmful levels, and the prevalence of anxiety (18%) was higher than the prevalence of depression (5.6%). Chi-square analysis revealed that more females than males were non-drinkers or non-hazardous drinkers, whereas more males reported drinking at hazardous and harmful levels (χ2 (3) = 33.92, P < 0.001). The levels of hazardous and harmful alcohol consumption detected by the AUDIT scores for men and women were higher than those reported in the general population (Degenhardt et al., 2000), however they are consistent with other studies, which point to a higher incidence of heavy drinking amongst users of Emergency Departments in Australia (Hulse et al., 2001; Roche et al., 2001; Stockwell et al., 2002).
. | Males (n = 452) . | Females (n = 360) . | Full sample (n = 812) . | |||
---|---|---|---|---|---|---|
AUDIT score*** (%) | ||||||
Non-drinker | 14.5 | 24.0 | 18.7 | |||
Non-hazardous | 46.8 | 55.8 | 50.8 | |||
Hazardous | 26.7 | 13.6 | 20.9 | |||
Harmful | 12.0 | 6.6 | 9.6 | |||
Tobacco consumption (%) | ||||||
Non-smoker | 55.5 | 62.8 | 58.7 | |||
Light smoker | 7.5 | 8.9 | 8.1 | |||
Moderate smoker | 16.8 | 14.7 | 15.9 | |||
Heavy smoker | 20.1 | 13.6 | 17.2 | |||
HADS anxiety score* (%) | ||||||
Normal | 72.2 | 63.1 | 68.2 | |||
Doubtful case | 11.7 | 16.5 | 13.8 | |||
Definite case | 16.1 | 20.4 | 18.0 | |||
HADS depression score (%) | ||||||
Normal | 85.9 | 90.5 | 87.9 | |||
Doubtful case | 7.2 | 5.6 | 6.5 | |||
Definite case | 7.0 | 3.9 | 5.6 |
. | Males (n = 452) . | Females (n = 360) . | Full sample (n = 812) . | |||
---|---|---|---|---|---|---|
AUDIT score*** (%) | ||||||
Non-drinker | 14.5 | 24.0 | 18.7 | |||
Non-hazardous | 46.8 | 55.8 | 50.8 | |||
Hazardous | 26.7 | 13.6 | 20.9 | |||
Harmful | 12.0 | 6.6 | 9.6 | |||
Tobacco consumption (%) | ||||||
Non-smoker | 55.5 | 62.8 | 58.7 | |||
Light smoker | 7.5 | 8.9 | 8.1 | |||
Moderate smoker | 16.8 | 14.7 | 15.9 | |||
Heavy smoker | 20.1 | 13.6 | 17.2 | |||
HADS anxiety score* (%) | ||||||
Normal | 72.2 | 63.1 | 68.2 | |||
Doubtful case | 11.7 | 16.5 | 13.8 | |||
Definite case | 16.1 | 20.4 | 18.0 | |||
HADS depression score (%) | ||||||
Normal | 85.9 | 90.5 | 87.9 | |||
Doubtful case | 7.2 | 5.6 | 6.5 | |||
Definite case | 7.0 | 3.9 | 5.6 |
P < 0.05
P < 0.001.
. | Males (n = 452) . | Females (n = 360) . | Full sample (n = 812) . | |||
---|---|---|---|---|---|---|
AUDIT score*** (%) | ||||||
Non-drinker | 14.5 | 24.0 | 18.7 | |||
Non-hazardous | 46.8 | 55.8 | 50.8 | |||
Hazardous | 26.7 | 13.6 | 20.9 | |||
Harmful | 12.0 | 6.6 | 9.6 | |||
Tobacco consumption (%) | ||||||
Non-smoker | 55.5 | 62.8 | 58.7 | |||
Light smoker | 7.5 | 8.9 | 8.1 | |||
Moderate smoker | 16.8 | 14.7 | 15.9 | |||
Heavy smoker | 20.1 | 13.6 | 17.2 | |||
HADS anxiety score* (%) | ||||||
Normal | 72.2 | 63.1 | 68.2 | |||
Doubtful case | 11.7 | 16.5 | 13.8 | |||
Definite case | 16.1 | 20.4 | 18.0 | |||
HADS depression score (%) | ||||||
Normal | 85.9 | 90.5 | 87.9 | |||
Doubtful case | 7.2 | 5.6 | 6.5 | |||
Definite case | 7.0 | 3.9 | 5.6 |
. | Males (n = 452) . | Females (n = 360) . | Full sample (n = 812) . | |||
---|---|---|---|---|---|---|
AUDIT score*** (%) | ||||||
Non-drinker | 14.5 | 24.0 | 18.7 | |||
Non-hazardous | 46.8 | 55.8 | 50.8 | |||
Hazardous | 26.7 | 13.6 | 20.9 | |||
Harmful | 12.0 | 6.6 | 9.6 | |||
Tobacco consumption (%) | ||||||
Non-smoker | 55.5 | 62.8 | 58.7 | |||
Light smoker | 7.5 | 8.9 | 8.1 | |||
Moderate smoker | 16.8 | 14.7 | 15.9 | |||
Heavy smoker | 20.1 | 13.6 | 17.2 | |||
HADS anxiety score* (%) | ||||||
Normal | 72.2 | 63.1 | 68.2 | |||
Doubtful case | 11.7 | 16.5 | 13.8 | |||
Definite case | 16.1 | 20.4 | 18.0 | |||
HADS depression score (%) | ||||||
Normal | 85.9 | 90.5 | 87.9 | |||
Doubtful case | 7.2 | 5.6 | 6.5 | |||
Definite case | 7.0 | 3.9 | 5.6 |
P < 0.05
P < 0.001.
More females than males reported zero tobacco consumption, but among those who did smoke, it was more often males who reported moderate or heavy consumption (χ2 (3) = 7.75, P = 0.052). Again, levels of tobacco consumption reported by participants in this study are higher than those detected in the Australian general population (Australian Institute of Health and Welfare, 2002), however they do reflect the higher incidence of smoking often reported amongst Emergency Department presentations (Richman et al., 1999; Horn et al., 2000, 2002).
Females more often reported borderline or clinical levels of anxiety (χ2 (2) = 7.73, P < 0.05). The levels of anxiety in this sample were higher than those in the general population (9.7%) and were greater amongst women than men (20.4% vs 16.1%). This gender difference is consistent, although to a much higher degree, with the higher levels of female anxiety in the general population (12.1% vs 7.1%) (Henderson et al., 2000).
The prevalence of depression in this Emergency Department (5.6%) was similar to that of the Australian general population (5.8%), but there was a higher incidence of depression among men (7.0%) than among women (3.9%). Although there was no significant association between gender and depression (χ2 (2) = 4.54, P > 0.05), the higher prevalence of depression amongst men is not consistent with national trends, where women report higher rates of depression (12%) than men (4.2%) (Henderson et al., 2000).
In order to investigate the relationship between (i) alcohol use and mental health, and (ii) tobacco use and mental health, eight one-way analyses of covariance (ANCOVA) were conducted, controlling for age, income, education and triage code, and conducted separately for male and female participants. The results of these analyses are presented below and illustrated in Figs 1–4.
Among males, anxiety scores differed significantly depending on AUDIT score, F(3,394) = 8.64, P < 0.001. Those in the harmful drinking category reported significantly more anxiety than those in other categories, F(1,394) = 15.43, P < 0.001, however non-drinkers also reported significantly more anxiety than either non-hazardous or hazardous drinkers, F(1,394) = 6.61, P < 0.05. There was no significant difference in the anxiety scores of non-hazardous and hazardous male drinkers, F(1,394) = 0.24, P > 0.05. Thus, for males the relationship between anxiety and alcohol consumption was described by a ‘U-shaped curve’, consistent with previous findings (Power et al., 1998; Rodgers et al., 2000b).
Similarly, there was a significant association between AUDIT category and HADS depression score among males, F(3,394) = 11.89, P < 0.001. Depression scores were higher in the harmful drinking category than in any other, F(1,394) = 18.28, P < 0.001, and higher in the non-drinking category than in either the non-hazardous or hazardous categories, F(1,394) = 11.75, P < 0.01. Non-hazardous and hazardous male drinkers did not differ in their depression scores, F(1,394) = 0.24, P > 0.05. Again, this pattern of results is consistent with the U-shaped curve, as illustrated in Fig. 1, and with previous studies (Power et al., 1998; Rodgers et al., 2000b).
Among females there was a significant association between AUDIT category and HADS anxiety score, F(3,306) = 3.32, P < 0.05. Females in the harmful drinking category reported significantly more anxiety than those in other categories [F(1,306) = 8.40, P < 0.01], however there was no significant difference between non-drinking females and those in the non-hazardous and hazardous categories, F(1,306) = 0.38, P > 0.05. Similarly, non-hazardous and hazardous female drinkers did not differ in their anxiety scores, F(1,306) = 0.13, P > 0.05. Thus, whereas for males the relationship between anxiety and alcohol consumption was described by a ‘U-shaped curve’, this was not the case for females.
A similar pattern described the association between alcohol consumption and depression for females, with those in different AUDIT categories receiving significantly different depression scores, F(3,306) = 4.60, P < 0.01. Females in the harmful drinking category reported more depression than those in other categories, F(1,306) = 11.32, P < 0.01, however there was no difference between either non-drinkers and non-hazardous or hazardous female drinkers, F(1,306) = 1.63, P > 0.05, or between non-hazardous and hazardous female drinkers, F(1,306) = 0.06, P > 0.05.
Among males, HADS anxiety score varied as a function of tobacco consumption, F(3,408) = 3.80, P < 0.05. Non-smoking males reported significantly less anxiety than those who did smoke, F(1,408) = 5.78, P < 0.05, however, levels of anxiety did not differ between light smokers and moderate or heavy smokers, F(1,408) = 2.32, P > 0.05, or between moderate and heavy smokers, F(1,408) = 0.01, P > 0.05.
By contrast, controlling for age, education, income and triage code, there was no association between HADS depression score and tobacco consumption among males F(3,408) = 1.83, P > 0.05 (Fig. 3).
Finally, whereas male non-smokers reported significantly less anxiety than those who did smoke, among females there was no relationship between HADS anxiety score and tobacco consumption [F(3,315) = 1.69, P > 0.05], and no relationship between HADS depression score and tobacco consumption [F(3,315) = 0.49, P > 0.05].
The same analyses were also run using only age as a covariate, and then adding the additional covariates (education, income, triage code) one at a time (tables not presented). The results of these additional analyses did not differ substantively from those presented here.
DISCUSSION
Findings from this study support previous evidence describing a nonlinear (‘U-shaped’) relationship between alcohol consumption and mental health among males (Power et al., 1998; Rodgers et al., 2000b). Our findings also suggest that the association between alcohol consumption and improved mental health among males may extend to levels of drinking labelled as ‘hazardous’ from a public health perspective. This runs counter to previous studies, which have found that symptoms of anxiety and depression are lower only amongst so-called ‘moderate’ drinkers (Rodgers et al., 2000a,b; Degenhardt and Hall, 2003).
In our sample the relationship between mental health and alcohol consumption was linear for females, despite a non-significant trend towards a ‘U-shaped’ relationship. This finding is inconsistent with previous studies (Degenhardt et al., 2000; Rodgers et al., 2000b) and brings into question the association between moderate alcohol consumption and better mental health among women. The finding is, however, consistent with the mortality literature, where this gender difference has been noted (Smith-Warner et al., 1998; Bovet and Paccaud, 2001). Although women's moderate consumption of alcohol is associated with positive health outcomes in relation to some diseases, it appears that the same consumption may be related to greater mortality with respect to other diseases (Smith-Warner et al., 1998; Bovet and Paccaud, 2001). Our findings suggest that this gender difference may extend to the area of mental health.
Population data point to a linear association between symptoms of anxiety and depression and smoking status (Australian Bureau of Statistics, 1998; World Health Organization, 2002). Findings from this study suggest that the same is true amongst patients presenting to an Emergency Department in Australia, although again only for males, and perhaps only with respect to anxiety. The apparent lack of association between tobacco consumption and mental health among females in our study raises questions about the link between smoking, anxiety and depression in women.
The present study has some notable limitations. First, while the study population is likely to be reasonably representative of Emergency Department patients, it is unclear to what extent the present findings can be generalized to the wider population. The demographic characteristics of this sample suggest an over-representation of people from lower socio-economic status groups, who consume tobacco and alcohol at higher rates than the Australian general population and experience a higher prevalence of some mental health problems (World Health Organization, 2002). Second, the screening tool used to assess mental health status (the HADS) is designed to detect clinically significant pathology—as such, it may underestimate the prevalence of less severe symptoms of anxiety and depression. Similar considerations apply to the AUDIT questionnaire, which may be more suited to detecting problematic use than lower levels of use. As with the HADS, the AUDIT is a self-report screening tool, and is, therefore, constrained by the truthfulness of the respondents' reports and the limitations of their memory. Finally, the HADS assesses state rather than trait anxiety and depression. Arguably, responses may be affected by the nature and severity of the respondent's clinical presentation. However, adjustment for urgency of case at presentation (triage code) did not appear to alter the nature of the associations between alcohol use and mental health status, suggesting that our results are unlikely to be biased by the severity of participants' clinical conditions.
Notwithstanding these limitations, the study has a number of strengths. Although not representative of the general Australian population, this sample represents well those presenting to public hospital Emergency Departments in Australia. It is also likely to be representative of individuals from low socio-economic status backgrounds, who tend to smoke and drink more heavily and experience a higher incidence of mental health problems, particularly anxiety and depression (World Health Organization, 2002; Claussen et al., 2003). Also, despite the use of screening instruments to measure alcohol consumption and mental health status, the observed patterns of nonlinear association between alcohol consumption and anxiety and depression are not dissimilar to those identified in other community surveys, at least as far as males are concerned (Degenhardt et al., 2000; Rodgers et al., 2000a,b).This is an important finding, as it has been replicated in a setting which lends itself to the application of intervention strategies.
CONCLUSIONS
This study directs attention to possible gender differences in patterns of association between alcohol and tobacco consumption and anxiety and depression. Among males, the observed relationship between alcohol consumption and mental health was non-linear, while for females it was linear. In addition, tobacco consumption was associated with greater anxiety among males, but not females. This study reiterates the need for further research attention to the mental health needs of non-drinkers, and to exploration of the differential relationship between alcohol use, tobacco use and mental health amongst males and females.
REFERENCES
ACEM (
Australian Bureau of Statistics (
Australian Institute of Health and Welfare (
Babor, T. F., De La Fuente, J. R., Saunders, J. and Grant, M. (
Barnett, N. P., Monti, P. M., Spirito, A., Colby, S. M., Rohsenow, D. J., Ruffolo, L. and Woolard, R. (
Bovet, P. and Paccaud, F. (
Caldwell, T. M., Rodgers, B., Jorm, A. F., Christensen, H., Jacomb, P. A., Korten, A. E. and Lynskey, M. T. (
Cassar, S., Hodgkiss, A., Ramirez, A. and Williams, D. (
Claussen, B., Davey Smith, G. and Thelle, D. (
Degenhardt, L. and Hall, W. (
Degenhardt, L. and Hall, W. (
Degenhardt, L., Hall, W. and Lynskey, M. (
Degenhardt, L., Hall, W., Teesson, M. and Michael, L. (
Ezzati, M., Lopez, A. D., Rodgers, A., Vander Orn, S. and Murray, C. J. L. (
Goldman, E. and Najman, J. M. (
Henderson, S., Andrews, G. and Hall, W. (
Horn, K., Gao, X., Williams, J., Helmkamp, J., Furbee, M. and Manley, W. (
Horn, K., Leontieva, L., Williams, J. M., Furbee, P. M., Helmkamp, J. C. and Manley, W. G. 3rd (
Hulse, G. K., Robertson, S. I. and Tait, R. J. (
Jorm, A. F., Rodgers, B., Jacomb, P. A., Christensen, H., Henderson, S. and Korten, A. E. (
National Health and Medical Research Council (
Power, C., Rodgers, B. and Hope, S. (
Richman, P. B., Dinowitz, S., Nashed, A., Eskin, B. and Cody, R. (
Roche, A. M., Watt, K., McClure, R., Purdie, D. M. and Green, D. (
Rockett, I. R., Putnam, S. L., Jia, H. and Smith, G. S. (
Rodgers, B., Korten, A. E., Jorm, A. F., Christensen, H., Henderson, S. and Jacomb, P. A. (
Rodgers, B., Korten, A. E., Jorm, A. F., Jacomb, P. A., Christensen, H. and Henderson, A. S. (
Smith-Warner, S. A., Spiegelman, D., Yaun, S. S. et al. (
Spirito, A., Barnett, N. P., Lewander, W., Colby, S. M., Rohsenow, D. J., Eaton, C. A. and Monti, P. M. (
Stockwell, T., McLeod, R., Stevens, M., Phillips, M., Webb, M. and Jelinek, G. (
Tait, R. J., Hulse, G. K., Robertson, S. I. and Sprivulis, P. (
Woolfenden, S., Dossetor, D. and Williams, K. (
World Health Organization (
Author notes
1Queensland Alcohol and Drug Research and Education Centre (QADREC), School of Population Health, The University of Queensland, Herston QLD 4006, Australia
2Violence Research Group, College of Medicine, University of Wales Health, South Wales, United Kingdom
3Accident and Emergency Department, Gold Coast Hospital, QLD 4215, Australia