Abstract
The rapid decline in cardiovascular deaths in Australia since 1970 has been attributed to a combination of improved medical intervention and a reduction in the prevalence of risk factors for cardiovascular disease. Medical advances, for example in the treatment of hypertension, have reduced the incidence of heart attack, while improvements in emergency care, rehabilitation and drug use have contributed to increased survival rates following heart attack [1–3]. At the same time, surveys by the National Heart Foundation [4] have shown that a reduction in the prevalence of risk factors, such as smoking and hypertension, is consistent with the decline in the incidence of cardiovascular disease. Despite this decline in risk factors, the Heart Foundation studies found that approximately 30% of Australians still have at least one modifiable risk factor [4]. Continued reductions in the prevalence of risk factors offer an opportunity for further improvements in cardiovascular health.
The reduction in risk factors for cardiovascular disease is largely the product of intensive campaigns to increase public awareness and change behaviour. For several decades people have been encouraged to quit smoking, exercise regularly and moderately, limit their alcohol consumption, reduce their salt and saturated fat intake and to maintain a healthy body weight. Public health messages, however, are more effective with some sections of the population than others. People from lower socioeconomic groups are roughly twice as likely to die from cardiovascular disease compared to those in the highest socioeconomic groups. Similarly, indigenous Australians are more than twice as likely to die from cardiovascular disease [5]. Although differential access to quality medical care cannot be discounted, the increased prevalence of risk factors among lower socioeconomic groups and indigenous Australians is a major contributor to this discrepancy in cardiovascular mortality.
Numerous studies show that people with mental illness are much more likely than other people to die from cardiovascular disease [6–9]. The only Australian data is provided by Ruschena et al. [10] who matched the Victorian Psychiatric Case Register with the register of deaths reported to the State Coroner. They found that unexpected deaths from natural causes, notably cardiovascular disease, was almost three times greater among people with schizophrenia (RR = 2.9, 95% CI = 2.1–3.9).
Apart from numerous studies into smoking, surprisingly little research has been done on the prevalence of cardiovascular risk factors among the mentally ill [11–13]. Kendrick [14] found that people with mental illness had high levels of cardiovascular and respiratory risk factors and symptoms, particularly smoking, obesity and hypertension. Similarly, Brown et al. [15] found that compared with the general population, people with schizophrenia ate a diet higher in fat and lower in fibre, were more likely to be overweight, and to smoke heavily but were less likely to drink alcohol.
The current study was designed to record the prevalence of risk factors for a range of physical health disorders among people with chronic mental illness. In this study, people with chronic mental illness are defined as those people attending community mental health services for the treatment of psychotic or affective disorders (e.g. schizophrenia, schizoaffective disorder, bipolar affective disorder) and who exhibit chronic psychiatric symptoms. This study examines a wider range of cardiovascular risk factors than previous studies [14, 15] and targets a vulnerable group not typically identified in large-scale studies of the Australian population such as the National Health Survey [16] or the Risk Factor Prevalence Study [4]. The results will enable comparisons to be made with the general population and provide a baseline against which future interventions can be evaluated.
Method
The methods used in this study have been described elsewhere [17]. Briefly, 234 outpatients of four Area Mental Health Services in the North-western Health Care Network Mental Health Program in Melbourne were interviewed about their health behaviour and underwent a brief physical examination. Participants were asked about a range of health behaviours including those related to preventive service use and HIV risk. This paper will report on the results related to cardiovascular risk.
Only patients judged by clinic staff to be capable of giving informed consent were approached for interview. The majority of participants were recruited into the study via their case manager or were approached at the clinic by the researcher who explained the aims and requirements of the study. A small number of participants were recruited when they contacted the researcher after seeing notices in the clinic advertising the project. Patients were excluded from participation if they were judged by their case manager to be too unwell; unable to understand the concept of informed consent; unable to speak English and an interpreter was not available; or were unwilling to participate.
The cardiovascular risk factors recorded in the study were cigarette smoking, exercise, alcohol consumption, salt intake, body mass index, blood pressure and history of hypercholesterolemia. Self-report items from the Risk Factor Prevalence Study [4] were used to obtain information on exercise, cholesterol screening and salt intake. The Standard Tobacco Questions developed by the Cardiovascular Behavioural Research Unit at the Anti-Cancer Council of Victoria were used to assess smoking habits. Blood pressure was taken using a recently calibrated electronic sphygmomanometer with standard cuffs. Two measures were taken 5 min apart and the average of the two readings was recorded. Participants were measured and weighed without shoes and wearing light street clothing. Participants were asked whether they had any previous diagnosis of hypercholesterolemia although blood lipids were not measured to validate self-report.
Data from the current study was compared with Australian population data provided by the Risk Factor Prevalence Study [4] (exercise, salt intake, blood pressure and history of hypercholesterolemia) and the 1995 National Health Survey [16] (alcohol, smoking and body mass index). Data was entered into SPSS (SPSS, Chicago, IL, USA) and descriptive analyses were conducted. Odds ratios (OR) and 95% confidence intervals (95% CI) were calculated to estimate the differences in the prevalence of risk factors between people with mental illness and the general Australian population.
Results
Sample
The sample consisted of 98 women and 136 men between the ages of 18–65 years who were living with chronic mental illness. Because a convenience sample was used it is difficult to accurately assess refusal rates. Approximately 20% of clients approached by the researcher refused to participate. However, it is possible that some patients refused via their case manager who did not subsequently inform the researcher.
Examination of records from the mental health services participating in the study indicates that the gender distribution of the sample was representative. Compared with the Risk Factor Prevalence Study [4] and the 1995 National Health Survey [16], participants in the current study were slightly overrepresented in the younger age groups and slightly underrepresented in the 40 years plus age groups.
The majority (79%) of the sample had a primary diagnosis of psychotic disorder, followed by major depressive disorder (9%), bipolar disorder (4%), personality disorder (4%) and other disorders (3%). More than half (51%) had been hospitalized for psychiatric illness in the last 12 months, and a further 17% had been hospitalized in the past 3 years. Twenty per cent of the sample were married or living in a de facto relationship at the time of interview and just over half (51%) lived with either their parents or partner and/or children. Few participants were currently employed and 92% received a government pension as their main form of income. Just over one-quarter (27%) were born in a non-English-speaking country. There were no Aboriginal or Torres Strait Islander respondents in the sample.
Smoking
Table 1 shows that 68% of men and 54% of women with mental illness were daily smokers. The majority (54%) were heavy smokers, smoking more than 20 cigarettes a day. Overall, participants were five times more likely to smoke than a community sample (OR = 5.2, 95% CI = 4.0–6.8). They were twice as likely to take up smoking (OR = 2.3, 95% CI = 1.8–3.0) and one-quarter as likely to quit (OR = 0.24, 95% CI = 0.16–0.39). In line with the rest of the population men were more likely to smoke than women. Although not statistically significant, participants with a diagnosis of a psychotic disorder were more likely to smoke than those with a non-psychotic diagnosis (OR = 1.8, 95% CI = 0.95–3.4).
The prevalence of cardiovascular risk factors for the mentally ill and for the general population
Body mass index
The majority of people with mental illness had a body mass index (kg/m2) over 25, which classifies them as overweight (Table 1). People with mental illness were more likely than the general population to be overweight (OR = 1.3, 95% CI = 1.0–1.7) and eight times more likely to be obese (OR = 7.8, 95% CI = 6.0–10.4). People who were obese were more likely to be women than men, to rate their health as fair to poor rather than good to excellent, to be non-smokers and to record a systolic blood pressure over 140 and a diastolic blood pressure over 80 (data not shown).
Medication and weight
Chi-squared analysis was performed to determine if the newer atypical antipsychotic medication was associated with increased weight among psychiatric patients. At the time of interview, 17% of participants were prescribed clozapine and 13% were prescribed olanzapine. Medication compliance was not biochemically validated, but analysis indicated that patients who were prescribed clozapine were almost four times more likely to be overweight or obese than those not prescribed clozapine (OR = 3.7, 95% CI = 1.4–10.0). No differences in body mass index were found between people prescribed olanzapine and those who were not.
Alcohol
Table 1 shows that people with mental illness were twice as likely to have abstained from alcohol in the past week than participants in the National Health Survey (OR = 2.0, 95% CI = 1.5–2.6). Almost 12% reported drinking at harmful levels as defined by the National Health and Medical Research Council, which was four times the rate of harmful drinking reported by a community sample (OR = 4.0, 95% CI = 2.7–6.0).
Exercise
Substantial proportions of people in the current study and the Risk Factor Prevalence Study [4] did no exercise in the 2 weeks preceding the interview (see Table 1). Participants were half as likely to undertake light exercise (OR = 0.54, 95% CI = 0.39–0.77) and one-quarter as likely to engage in vigorous exercise (OR = 0.26, 95% CI = 0.17–0.39). The mentally ill sample were somewhat more likely to report that they had walked for exercise in the last 2 weeks.
Blood pressure
Few differences in blood pressure were found between the study group and other people. Mentally ill women were more likely to record systolic blood pressure above 140 than other women, however, both men and women with mental illness were slightly less likely to record diastolic blood pressure over 80.
Salt intake
One-fifth (22%) of people with mental illness reported that they usually added salt to their food after it was cooked compared with 14% of people surveyed in the Risk Factor Prevalence Study [4] (OR = 1.7, 95% CI = 1.2–2.4). More than half (56%) of the general population reported never adding salt to their food after it has been cooked compared with 48% of the mentally ill.
History of hypercholesterolemia
The majority of both the mentally ill and the general population have never had a blood cholesterol test. Sixty-four per cent of men with mental illness were not aware of their cholesterol status compared with 53% of men in the general population. However, women with mental illness were more likely to have had their blood cholesterol measured than women in the general population (48% cf. 38%). Among the mentally ill sample, 12% of men and 14% of women reported that they had been told by a health professional that they had high blood cholesterol. This compared with 17% of men and 11% of women in the general population who had been told they had high cholesterol.
Discussion
Economic and social disadvantage are additional impediments to those with mental illness achieving control over their lives. Psychosocial factors may be linked to coronary heart disease in three ways: healthrelated behaviours, pathophysiological changes and access and content of medical care [18]. This study provides evidence that people with mental illness have poor health-related behaviour and receive inadequate medical or public health attention directed at altering these factors.
Compared with a community sample, people with mental illness recorded a significantly higher prevalence on five of the seven cardiovascular risk factors examined in this study. They were much more likely to smoke, to be overweight or obese, to not exercise and to add salt to their food. The U-shaped prevalence curve in alcohol consumption indicates that not only are the mentally ill more likely to drink at harmful levels, but because they are more likely to abstain they are less likely to gain the protective benefits of consuming small amounts of alcohol [19]. The one risk factor for which the mentally ill did not record higher levels than the general population was for blood pressure.
Results relating to blood cholesterol are more difficult to interpret because cholesterol screening rather than blood lipids was measured. Consistent with ascertainment due to having been tested, more men in the general population and more women with mental illness had been told they had high cholesterol. Although it is impossible to know the exact relationship between cholesterol screening and blood lipids, the prevalence of overweight and obesity among the study group suggests that the mentally ill are at considerable risk for hypercholesterolemia. Rates of cholesterol screening should ideally reflect this by being much higher than they were.
The prevalence of smoking and obesity are the most worrying results from the survey. Smoking prevalence recorded in this study is consistent with a recent survey of nearly 1000 Australians living with a psychotic illness which reported smoking rates of 73% for men and 56% for women [20]. Similarly, a study by Kelly and McCreadie [21] found that 71% of men and 42% of women with mental illness were smokers. Studies of psychiatric inpatients have found an even higher prevalence of smoking [11, 22], however, studies with broader sampling methods, including the current study, suggest that rates of between 65% to 75% for men, and between 40% and 55% for women, are better estimates of the prevalence of smoking among people with chronic mental illness.
How the prevalence of smoking among the mentally ill has changed over time is undocumented. In the general male population the prevalence of smoking has fallen from 72% in the 1940s to its current 27% [23]. In contrast, smoking rates of mentally ill men currently hovers at 68%, while for women it is 54%. It is unclear whether smoking rates for mentally ill men have remained consistent or have increased or decreased since the 1940s. This uncertainty highlights the need for repeat crosssectional studies to monitor the prevalence of cardiovascular risk factors among the mentally ill and to gain a better understanding of how this population responds to public health campaigns.
Consistent with previous studies, the current study found that the mentally ill had a far higher prevalence of being overweight or obese compared with the general population and that women were more likely to be overweight or obese than men [22, 24]. A recent nutritional assessment study of 38 mentally ill residents in supported accommodation in Sydney found that 37% were obese and 34% were overweight [24]. Milson and Avard [25] found that the diet of the mentally ill is characterized by a high intake of fast food and a general lack of understanding of dietary principles, both of which are amenable to intervention. It has further been argued that the bulk of the weight gain among people with chronic mental illness occurs in the early stages of the illness and therefore early, rather than late, intervention is likely to be most effective [22].
Less amenable to intervention may the weight gain associated with some of the newer atypical antipsychotic medication. This study found that respondents prescribed clozapine were significantly heavier than patients on other medications. However, in contrast with several other studies, respondents prescribed olanzapine were not heavier than other respondents [26, 27]. The current results should be interpreted with considerable caution as this study was not specifically designed to examine the weight gain effect of medication. Specifically, patients’ weight prior to taking clozapine or olanzapine, the doses prescribed and compliance with medication was not recorded. Despite these cautions, if the association between the newer atypical antipsychotics and weight gain holds true, then the benefits of the newer drugs may need to be balanced against the possible increase in risk for cardiovascular disease.
Despite suggestions that practitioners need to consider methods for minimizing the impact of weight gain induced by antipsychotic drugs [26], the low levels of light or vigorous exercise among the mentally ill indicates that exercise has not been adequately promoted. The higher rates of walking among the sample is probably due to the higher proportion of people with mental illness who do not drive or own cars.
The mixed results on alcohol consumption indicate that clinic-wide interventions targeting a reduction in alcohol consumption would not be appropriate. Rather, individuals at risk of drinking at harmful levels should be identified and targeted by their case workers. The establishment of dual diagnosis units in some mental health services to specialize in treating psychiatric patients with a comorbid diagnosis of substance abuse indicates that this approach is currently being implemented.
The high prevalence of obesity among the mentally ill stands in contrast to the non-significant differences in blood pressure between the study group and the general population. One explanation may be that the large sphygmomanometer cuff was exceptionally large and too big for all but two participants. It is possible that inappropriately small cuffs were used on some individuals which may have contributed to lower blood pressure readings. Alternatively, many of the medications used by this group cause mild hypotension as a side effect. Finally, compared with the measures required to reduce smoking and obesity, prescription and monitoring of antihypertensive medication is a relatively simple intervention and may have been already prevalent in the study group.
A limitation of this study is that comparison data from the Risk Factor Prevalence Study [4] is now a decade old which raises several issues. The prevalence data from the Risk Factor Prevalence Study over the 1980s showed continual improvement in all risk factors except for body fatness and blood lipids. It is reasonable to assume that this trend has continued since 1989 and therefore that the differences in the prevalence of risk factors between the mentally ill and the general population in 1999 are underestimated.
Although the ethnic mix in the current study is similar to that of the census, the 10 years age difference between the study sample and the reference group may mean that the ethnic mix may have changed, increasing or decreasing the proportion of certain ethnic groups who are known to have cardiovascular risk or protective benefits. The Risk Factor Prevalence Study did not report the country of birth of their sample and it is difficult to say what effect, if any, differential migration patterns may have on the comparisons between the mentally ill sample and the reference group.
Limitations of the data did not allow socioeconomic status (SES), a determinant of risk-factor status, to be controlled. While it is noted that the low income of this group is likely to contribute to their risk-factor profile, low SES is also a relatively stable characteristic of the chronically mentally ill population and is an essential consideration when determining intervention programs.
Furthermore, the higher prevalence of risk factors in the mentally ill sample compared with a community sample needs to be interpreted in light of the larger proportion of people in the younger age groups in the current study. Cardiovascular risk factors tend to increase with age, which suggests that if the two groups were fully matched for age and included equal proportions of older people then the discrepancy between the two groups would again be larger than the current results suggest.
People with chronic mental illness have more contact with health professionals than other people, yet the results of this study clearly show that they have a significantly higher prevalence of key risk factors for cardiovascular disease. It may be argued that health professionals have neglected the health needs of the mentally ill and that this neglect contributes to the excess of cardiovascular mortality in this group. Interventions aimed at reducing the prevalence of risk factors for cardiovascular disease among the mentally ill are urgently needed. Psychiatrists, general practitioners and public health experts need to work together towards reducing cardiovascular risk and to prevent the gap in health status between those with mental illness and those without from widening even further.
The inclusion of detailed information on physical risk factors such as smoking, obesity, alcohol abuse and illicit drug use in the initial assessment, and individual service plans for patients managed by community mental health services, and additional training for community mental health service case managers on the physical health of their clients are obvious first steps in improving the physical health of the mentally ill.
In addition, individuals at high risk for cardiovascular and respiratory disease should receive tailored interventions such as referral to dietitians, exercise programs, Quit programs and other health promotion interventions.
Finally, studies are urgently needed which trial and evaluate interventions to effectively modify risk factors in this group.
Footnotes
Acknowledgements
This study was funded by a project grant from the Victorian Health Promotion Foundation. The authors are grateful for the support of the managers of the Inner West Area Mental Health Service (AMHS), the Northwest AMHS, the Mid-west AMHS and the South-west AMHS of the North-west Health Care Network Mental Health Program.