Psychological distress is associated with a range of high-priority health conditions affecting working Australians
Libby Holden, School of Medicine, Griffith University, University Drive, Meadowbrook, QLD 4131. Fax: (07) 3382 1338; e-mail: email@example.com
Background: Psychological distress is growing in prevalence in Australia. Comorbid psychological distress and/or depressive symptoms are often associated with poorer health, higher healthcare utilisation and decreased adherence to medical treatments.
Methods: The Australian Work Outcomes Research Cost-benefit (WORC) study cross-sectional screening dataset was used to explore the association between psychological distress and a range of health conditions in a sample of approximately 78,000 working Australians. The study uses the World Health Organization Health and Productivity Questionnaire (HPQ), to identify self-reported health status. Within the HPQ is the Kessler 6 (K6), a six-item scale of psychological distress which strongly discriminates between those with and without a mental disorder. Potential confounders of age, sex, marital status, number of children, education level and annual income were included in multivariate logistic regression models.
Results: Psychological distress was significantly associated with all investigated health conditions in both crude and adjusted estimates. The conditions with the strongest adjusted association were, in order from highest: drug and alcohol problems, fatigue, migraine, CVD, COPD, injury and obesity.
Conclusions: Psychological distress is strongly associated with all 14 health conditions or risk factors investigated in this study. Comorbid psychological distress is a growing public health issue affecting Australian workers.
Psychological distress refers to subjective emotional distress1 and as the level of distress increases, there is an increasing likelihood of the person having symptoms sufficient to meet diagnostic criteria for a mental disorder.1,2 Depressive symptoms and/or psychological distress are often comorbid with other chronic diseases and can lead to poorer health outcomes.1–9 When psychological distress is comorbid with a physical health condition, poorer quality of life,10–12 higher healthcare usage,10,13 decreased adherence to medical treatments,6,14–16 and significantly increased disability days6,10,17–19 have been reported.
Several reports by the Australian Institute of Health and Welfare (AIHW) have identified mental health as a priority health condition.20–23 The Australia's Health 2008 report highlights a near two-fold increase in the prevalence of very high psychological distress from 1997 to 2005.24 It also reports that comorbidity involving more than one mental illness or having at least one mental illness with one or more physical illnesses is common among those with mental illness.24 Although mental health disorders are found to be increasingly prevalent in Australia, only 35% of Australians with a 12-month mental health disorder in 2007 accessed mental health services.25 Thus, in Australia, mental health disorders are both prevalent and under treated.
A strong association has been found between impaired mental health, particularly depression, and other health conditions;3,4,21,26,27 between depression or psychological distress and workplace productivity loss;28–33 and between depression and injury.34–36 An Australian study found that one of the strongest correlates for current depression was having a medical condition.37 Other studies have found that depressed people report significantly more comorbid medical conditions than non-depressed people.38,39 A World Health Organization (WHO) survey of 60 countries found that depression that is comorbid with physical health conditions incrementally worsens health compared to depression alone, compared with any of the chronic diseases alone, and compared with any combination of chronic diseases without depression.4
The prevalence of psychological distress has been found to increase with multi-morbidity (the presence of more than one health condition) after accounting for disease severity.1,2,40 One study demonstrated that the risk of psychological distress increased by a factor of five for those with high level multi-morbidity compared to those without multi-morbidity.40 Patients with multi-morbidity in general practice represent the rule rather than the exception.2,41,42 It is estimated that the number of Americans with multi-morbidity will increase from 60 million in 2000 to 81 million by 2020.2
Although there is a growing body of international evidence of psychological distress being associated with a range of comorbid health conditions, few studies are available that demonstrate these associations within an Australian context.37,41,43 Of the three studies found, the National Survey of Mental Health and Wellbeing, reported the strongest correlates for major depression as being unemployed, smoking and having a medical condition.37 The others reported psychological distress to be among one of the highest comorbidities presenting in general practice.41,43
Employee health and wellbeing is of increasing interest to employers as health status is related to productivity.32,44,45 To date there has been no Australian employee-based study on the comorbidity of psychological distress in relation to physical health disorders.
Using a large sample of working Australians, this study ascertains the extent to which psychological distress is comorbid with other health conditions. This information will not only guide occupational health and safety practice, but will inform medical clinicians of the health conditions most strongly associated with psychological distress that may warrant further investigation.
The Australian Work Outcomes Research Cost-benefit (WORC) project cross-sectional data set of approximately 78,000 working Australians was used to explore associations between psychological distress and a range of self-reported health conditions. Data were obtained using a validated self-reported written questionnaire.
Employees of 58 large Australian-based companies were invited to participate in the WORC study. The surveys were undertaken between October 2004 and December 2005. The response rate was 24.7% (90,279 employees). From this, 78,587 valid surveys were obtained for this sub-study. Respondents' demographic characteristics available through the HPQ are summarised in Table 1. Those employees aged less than 18 years and more than 70 years (0.2%) were excluded from the study, as these age groups are not usually in the workforce. The study sample has more females than the Australian workforce (65% and 45%, respectively).46 It also has greater representation of workers from industries of health, education and government administration; and fewer from retail, construction and mining.46 The average income and education level are fairly representative of the Australian population.46 Generalisation of these findings to all working Australians should be done with caution, particularly for blue-collar workers.
Table 1. Results of bi-variate logistic regression of demographic characteristics and psychological distress status.
|Agea||77,823|| || || || || |
| 18-29 years|| ||17||21||1.00|| || |
| 30-44 years|| ||43||46||0.86||0.78-0.95||0.004|
| 45-59 years|| ||37||32||0.69||0.62-0.77||<0.001|
| 60-70 years|| ||3||2||0.41||0.30-0.56||<0.001|
|Sex||77,841|| || || || || |
| Female|| ||65||35||1.00|| || |
| Male|| ||67||33||0.91||0.84-0.99||0.02|
|Marital status||77,626|| || || || || |
| Separated, divorced, widowed, never married|| ||28||40||1.00|| || |
| Married or cohabitating|| ||72||60||0.59||0.54-0.64||<0.001|
|Number of children||77,620|| || || || || |
| Nil|| ||70||67||1.00|| || |
| 1-3 children|| ||28||30||1.13||1.04-1.23||0.005|
| 4 or more children|| ||3||3||1.23||0.98-1.54||0.08|
|Education level||77,421|| || || || || |
| Completed college or university|| ||48||42||1.00|| || |
| Some college|| ||27||30||1.27||1.16-1.40||<0.001|
| Completed high school|| ||10||13||1.44||1.27-1.63||<0.001|
| Did not complete high school|| ||14||16||1.27||1.13-1.42||<0.001|
|Annual wageb||76,208|| || || || || |
| ≥$100,00 per year|| ||7||5||1.00|| || |
| $75,000-99,999 per year|| ||10||8||1.12||0.89-1.41||0.333|
| $50,000-74,999 per year|| ||36||34||1.37||1.13-1.66||0.001|
| $40,000-49,999 per year|| ||20||26||1.83||1.50-2.22||<0.001|
| $30,000-39,999 per year|| ||14||16||1.65||1.34-2.03||<0.001|
| ≤$29,999 per year|| ||13||13||1.38||1.12-1.71||0.003|
Participants completed the validated World Health Organization's Health and Productivity Questionnaire (HPQ).45 The HPQ identifies self-reported health status for a range of health conditions.45 In this study, for each self-reported health condition, health status has been coded as ‘yes’ if respondents reported having the condition and either currently or previously having professional treatment, and ‘no’ if they reported never having had the condition. Respondents were excluded if they reported having had the condition but never having received treatment as respondents may have incorrectly self-diagnosed the condition (average of 0.05% excluded per condition). The following health conditions were included in the analyses as these were available in the HPQ and identified as Australian health priorities (or risk factors) or conditions found in the literature to affect productivity: arthritis, asthma, back/neck pain, cancers, chronic obstructive pulmonary disease (COPD), cardiovascular disease (CVD), psychological distress, drug and alcohol problems (D&A), diabetes, fatigue, high blood pressure, high cholesterol, injury, migraine and obesity.
Embedded in the HPQ is the Kessler 6 (K6), a six-item scale of psychological distress with excellent internal consistency, and reliability which strongly discriminates between community mental health cases and non-cases.47 Each of the six items on the K6 has a 0-4 scale with zero being ‘none of the time’ and four being ‘all of the time’; and a total range of 0-24.48 The optimal cut point on the total K6 score to equalise false-positive and false-negative results is 13.48 We have used a dichotomous variable to indicate ‘yes’ or ‘no’ for the presence of high psychological distress where a score of 0-12 is coded as ‘no’ and a score of 13+ is coded as ‘yes’. The maximum score on the K6 is 24. Due to the design of the HPQ, respondents with answer bias that record a response “Not at all” to physical health symptom questions continue to the K6 questions and answer the maximum score “All of the time” on the K6. Therefore, the K6 = 24 group consists of those that have severe psychological distress and those that do not read the questions. As it is not possible to guarantee that all answer bias respondents have been removed from the sample, instances where the K6 = 24 have been removed from analyses (<1% excluded).
Univariate logistic regression was used to explore the crude association between psychological distress and each health condition or potential confounder. Multivariate logistic regression was used to explore health conditions associated with psychological distress; adjusting for age, sex, marital status, number of children, education level and median annual wage as a proxy for socio-economic status. Separate models were undertaken for each health condition to avoid co-linearity. Backward stepwise methods were used to exclude non-significant confounders from each model for each health condition if p≥0.05. Sex and/or education were significant in some models but removed from other models if not significant at p<0.05. Odds ratios (OR) and 95% confidence intervals (95%CI) are reported for all variables retained. The odds ratio is the estimate of odds of a certain event (e.g. disease) occurring in the exposed group compared to the unexposed group; in this case for those with psychological distress compared to those without.
A sample of 77,841 workers was analysed. Demographic characteristics are described in Table 1. Of the sample, 65% were female and 35% male. The two largest age groups were aged 30-44 years and 45-59 years, comprising 80% of the sample, with only 17% aged less than 30 and 3% aged over 60 years. Seventy-one per cent were married or cohabiting. Nearly 70% had no children. Nearly half had completed a tertiary qualification (48%). Approximately half earned $50,000 or more per year.
The unadjusted ORs of having comorbid psychological distress with health conditions are listed in Table 2. The conditions with the highest likelihood were: D&A (OR = 14.0), fatigue (OR = 5.0), migraine (OR = 2.8), COPD (OR = 2.5), CVD (OR = 2.2), and to a lesser extent injury (OR = 2.0), obesity (OR = 1.9), back/neck pain (OR = 1.9), diabetes (OR = 1.7), arthritis (OR = 1.5), asthma (OR = 1.5), cancer (OR = 1.4), high blood pressure (OR = 1.3) and high cholesterol (OR = 1.3).
Table 2. Unadjusted health conditions associated with psychological distress.
|Arthritisa||67,085|| || || || |
| No arthritis|| ||96||1.00|| || |
| Arthritis|| ||4||1.48||1.22-1.80||<0.001|
|Asthmaa||67,935|| || || || |
| No asthma|| ||94||1.00|| || |
| Asthma|| ||6||1.47||1.26-1.70||<0.001|
|Back / neck paina||71,949|| || || || |
| No back / neck pain|| ||70||1.00|| || |
| Back / neck pain|| ||30||1.93||1.77-2.10||<0.001|
|Cancers (not skin)a||77,183|| || || || |
|No cancers|| ||97||1.00|| || |
|Cardiovascular diseasea||76,709|| || || || |
|No cardiovascular disease|| ||99||1.00|| || |
|Cardiovascular disease|| ||1||2.18||1.59-2.99||<0.001|
|COPD / bronchitis / emphysemaa||74,288|| || || || |
|No COPD / bronchitis / emphysema|| ||99.6||1.00|| || |
|COPD / bronchitis / emphysema|| ||0.4||2.46||1.61-3.77||<0.001|
|Diabetesa||76,489|| || || || |
|No diabetes|| ||98||1.00|| || |
|Drug and alcohol problems||74,656|| || || || |
| No drug and alcohol problems|| ||99.8||1.00|| || |
| Drug and alcohol problems|| ||0.2||14.00||9.75-20.10||<0.001|
|Fatigue / sleep problemsa||77,322|| || || || |
| No fatigue / sleep problems|| ||99||1.00|| || |
| Fatigue / sleep problems|| ||1||5.04||3.94-6.44||<0.001|
|High blood pressurea||73,259|| || || || |
| No high blood pressure|| ||92||1.00|| || |
| High blood pressure|| ||8||1.31||1.14-1.49||<0.001|
|High cholesterola||70,932|| || || || |
| No high cholesterol|| ||94||1.00|| || |
| Cholesterol|| ||6||1.27||1.08-1.50||0.003|
|Migraine / severe headachea||72,836|| || || || |
| No migraine / severe headaches|| ||90||1.00|| || |
| Migraine / severe headaches|| ||10||2.80||2.54-3.10||<0.001|
|Obesitya||73,939|| || || || |
| No obesity|| ||89||1.00|| || |
| Obesity|| ||11||1.94||1.75-2.15||<0.001|
|Workplace injurya||77,490|| || || || |
| No injury|| ||93||1.00|| || |
| Injury|| ||7||1.96||1.75-2.21||<0.001|
|Comorbidity||77,841|| || || || |
| No health conditions|| ||13||1.00|| || |
| 1 health condition only|| ||16||1.23||0.90-1.68||0.19|
| 2-4 health conditions|| ||41||2.87||2.22-3.71||<0.001|
| 5-7 health conditions|| ||19||8.75||6.80-11.26||<0.001|
| 8-10 health conditions|| ||8||18.88||14.63-24.37||<0.001|
| 11 or more health conditions|| ||3||33.81||26.00-43.97||<0.001|
In adjusted models (adjusting for socio-demographic characteristics), all investigated health conditions had a significant association with psychological distress. The conditions with the greatest association were: D&A, fatigue, migraine, CVD, COPD, and to a lesser extent back/neck pain, injury, obesity, diabetes, arthritis, high blood pressure, high cholesterol, asthma and cancer (Table 3). The likelihood of having psychological distress also increased as the number of health comorbidities increased.
Table 3. Adjusted health conditions associated with psychological distress (adjusting for age, marital status, number of children, sex, education level, and annual income).
|Arthritisa||65,230|| || || || |
|No arthritis|| ||96||1.00|| || |
|Asthmaa||65,937|| || || || |
|No asthma|| ||94||1.00|| || |
|Back / neck paina||69,898|| || || || |
|No back / neck pain|| ||70||1.00|| || |
|Back / neck pain|| ||30||1.98||1.81-2.16||<0.001|
|Cancers (not skin)a||74,970|| || || || |
|No cancers|| ||97||1.00|| || |
|Cardiovascular diseasea||74,506|| || || || |
|No cardiovascular disease|| ||99||1.000|| || |
|Cardiovascular disease|| ||1||2.58||1.85-3.60||<0.001|
|COPD / bronchitis / emphysemaa||72,150|| || || || |
|No COPD / bronchitis / emphysema|| ||99.6||1.00|| || |
|COPD / bronchitis / emphysema|| ||0.4||2.43||1.57-3.78||<0.001|
|Diabetesa||74,295|| || || || |
|No diabetes|| ||98||1.00|| || |
|Drug and alcohol problems||72,490|| || || || |
|No drug and alcohol problems|| ||99.8||1.00|| || |
|Drug and alcohol problems|| ||0.2||12.80||8.82-18.57||<0.001|
|Fatigue / sleep problemsa||75,103|| || || || |
|No fatigue / sleep problems|| ||99||1.00|| || |
|Fatigue / sleep problems|| ||1||5.14||3.98-6.64||<0.001|
|High blood pressurea||71,153|| || || || |
|No high blood pressure|| ||92||1.00|| || |
|High blood pressure|| ||8||1.58||1.36-1.82||<0.001|
|High cholesterola||68,890|| || || || |
|No high cholesterol|| ||94||1.00|| || |
|Migraine / severe headachea||70,735|| || || || |
|No migraine / severe headaches|| ||90||1.00|| || |
|Migraine / severe headaches|| ||10||2.67||2.41-2.96||<0.001|
|Obesitya||71,883|| || || || |
|No obesity|| ||89||1.00|| || |
|Workplace injurya||75,281|| || || || |
|No injury|| ||93||1.00|| || |
|Comorbidity||75,891|| || || || |
|No health conditions|| ||13||1.00|| || |
|1 health condition only|| ||16||1.39||1.01-1.92||0.04|
|2-4 health conditions|| ||41||3.36||2.57-4.38||<0.001|
|5-7 health conditions|| ||19||10.94||8.40-14.23||<0.001|
|8-10 health conditions|| ||8||24.10||18.45-31.49||<0.001|
|11 or more health conditions|| ||3||46.39||35.20-61.14||<0.001|
Associations between factors were determined in this cross-sectional study; however, no conclusions can be drawn regarding causal pathways. After accounting for socio-demographic details, a significant association was found between psychological distress and all investigated health conditions in unadjusted estimates. The strength of that association varied between conditions. The health conditions with the strongest adjusted associations with psychological distress were: drug and alcohol problems, fatigue, migraine, CVD, COPD, injury, obesity, diabetes, arthritis, high blood pressure and high cholesterol, asthma and cancer.
These finding are supported by international studies1,3,26,27,38,39,49,50 and to some extent in Australia.37,41,43 One Australian study found that one of the strongest correlates for reported current depression was having a medical condition.37 Similarly, Britt43 and Saltman41 also demonstrate high prevalence of comorbid psychological distress in Australian general practice. Overseas research has found that depressed people report significantly more comorbid medical conditions,38,39 with one Netherlands study reporting an extensive range of 26 disease categories associated with depression.3
Our model adjusts for known demographic characteristics, an approach supported by Marmot's social gradient theory, which highlights the socio-economic determinants of health status.51 More recently, Manoux and Marmot (2005) argued that social class had a powerful influence on psychosocial vulnerability, as individuals in lower social strata have fewer psychosocial resources to cope with life events.52 Our findings highlight that these associations exist, but are not causally explained by purely social gradient. It could be that work-related factors may be another causal link influencing the relationship between health status and psychological distress in our study subjects. Rather, aspects of employment such as structure, status, income, high demands, low decision authority, effort-reward imbalance and social interaction play an important role in mental health.53 The nature of the relationship between employment status and mental illness is unclear as to both direction of causality and the impact of other confounding factors. However, regardless of the causal pathway and confounding factors, the presence of this relationship between psychological distress and other chronic health conditions warrants further public health consideration.
The findings that so many health conditions reported in the present study were significantly associated with an elevated risk for psychological distress and prior evidence that the combination of psychological distress, and chronic physical disease can lead to poorer health outcomes,1,4,7 has a wide-ranging impact on health promotion and screening programs and general practice. Given that the majority of individuals have at least one physical health condition, the high prevalence of mental disorders,1,4,6,21 the low treatment rates for mental disorders,54 and the fact that mental disorder (and high level psychological distress) is costly and disabling,38,54–59 the Australian Government's recent focus on preventative healthcare60 should include providing adequate screening and early intervention of mental health problems in working Australians.
This study also found that the number of comorbidities was strongly associated with psychological distress. There is Australian41,43 and overseas1,2 evidence to support this association between psychological distress and multi-morbidity. Psychological distress has been found to increase with multi-morbidity when disease severity is accounted for.1,2,40 Patients with multi-morbidity in general practice represent the rule rather than the exception.2,41,42 These rates are likely to increase with the ageing population, the increasing prevalence of chronic disease, and the improvements in medical technologies that enable people with chronic illness to live longer.61,62 Given the strong likelihood of someone with multiple chronic conditions having comorbid psychological distress, as evidenced by our findings, psychological screening of patients who present with multi-morbidity should be advocated in general practice.
The limitations of this study need to be considered with these findings. Participation of employees in the WORC study was voluntary; hence, selection bias may be present in this sample. However, a prior study has shown that respondents and non-respondents to a HRA survey have comparable levels of depression.63 The response rate of 24.9% may be considered low by epidemiologic standards; however, the response rate is typical of the response rates achieved by employee administered health questionnaires in large organisations.64 There is also potential bias associated with the self-reported nature of data regarding health conditions.
In addition, we have previously shown in regression analysis of response rate of the employers versus prevalence of psychological distress, there was no effect of response rate on prevalence estimates.33 The sample has some characteristics that are not indicative of the general working population. The sample has a higher proportion of females, people aged under 45 years, and people with a higher education level. Given that people from lower socio-economic backgrounds tend to have poorer health and health deteriorates with age, it is likely that these results are underestimating the true extent of the problem.
Our findings highlight the importance of comorbid psychological distress as a public health issue for working Australians, as demonstrated by its strong association with all health conditions explored. These results, combined with a two-fold increase in psychological distress in Australia from 1997 to 2005,24 the low rates of treatment-seeking,33,54 and reported poorer health outcomes for people with comorbid psychological distress or depression1–9 indicate comorbid psychological distress as a growing challenge for Australia.
Psychological distress is strongly associated with all 14 health conditions or risk factors investigated in this study. Comorbid psychological distress is a growing public health issue affecting Australian workers. Early detection and intervention is needed to address this growing problem.
The authors would like to acknowledge the contribution of Professor Ron Kessler and Dr Philip Wang for their guidance in establishing the study protocol. This work was financially supported by, 1) the Department of Health and Ageing, Mental Health Strategy Branch, Australian Government, Canberra, ACT; 2) beyondblue: the national depression initiative, Melbourne, VIC, Australia; and 3) The Australian Rotary Health Research Fund, Parramatta, NSW, Australia.