The Global Burden of Disease study (GBD), now in its second iteration, has been and will continue to be very influential in shaping health priorities. By providing empirical estimates of disease burden based on robust methodologies the GBD has changed government thinking about priorities for health service delivery, and provides an empirical basis for setting priorities for service reform, funding allocations and research priorities based on evidence of burden and potential impact.
However, as highlighted by Charlson et al. (Reference Charlson, Baxter, Dua, Degenhardt, Whiteford and Vos2015) the GBD presentation does not always tell the full story. In any accounting exercise, decisions need to be made about classifications and the choice of how to summarise a wealth of data into the most important headline indicators. The GBD has focused on years of life lost (YLLs) and years lived with disability (YLDs; Whiteford et al. (Reference Whiteford, Degenhardt, Rehm, Baxter, Ferrari, Erskine, Charlson, Norman, Flaxman, Johns, Burstein, Murray and Vos2013). Charlson et al. (Reference Charlson, Baxter, Dua, Degenhardt, Whiteford and Vos2015) provide illuminating additional information about the burden of mortality due to mental illness than could be shown in the YLL summary measures presented in the main GBD reports. These data show that although YLLs directly attributed to mental illness are a cause of concern, they are small in comparison to the excess mortality observed in people with mental illness using natural history models. This gap could be even larger than Charlson et al. have calculated. Schizophrenia is the only mental disorder currently included in the cause-specific calculation of YLLs. The natural history models additionally examine excess mortality in major depressive disorder, bipolar disorder and substance use disorders, but anxiety disorders have not been included. Our own data from Western Australia (WA) found that anxiety disorders have high prevalence and at least some anxiety disorders are associated with similar levels of excess mortality as depression (Lawrence et al.
Reference Lawrence, Hancock and Kisely2013). More epidemiological research is needed on the health outcomes of people with anxiety disorders before they could be included in this analysis, but if the WA results are confirmed in other places the burden of excess mortality associated with mental illness could be even higher than as stated by Charlson et al.
The high level of excess mortality associated with mental illness is one indicator of disturbing inequalities in health outcomes for people with mental illness that deserve serious attention (Thornicroft, Reference Thornicroft2013). The gap between the YLL analysis and the results of the natural history models is instructive in what it implies about the nature of the gap in life expectancy for people with mental illness and the possibility that it could be reduced even without a major new breakthrough in treating or preventing mental illness. With the most notable exception of dementia, which remains an incurable, terminal illness, there are no known biological mechanisms to believe that mental illnesses are direct causes of death. This gives hope that the excess mortality in people with mental illness is an inequality that is potentially remediable even while we do not yet know how to prevent or cure all mental illness. For instance, there are a range of evidence-based treatments for cardiovascular disease (CVD) including procedural interventions such as angioplasty, stenting and bypass grafts and pharmaceutical interventions such as aspirin, angiotensin-converting enzyme inhibitors and beta-blockers. However, people with both mental illness and CVD are less likely to receive these interventions than people with CVD who do not have mental illness (Druss et al.
Reference Druss, Bradford, Rosenheck, Radford and Krumholz2000; Mitchell et al.
Reference Mitchell, Malone and Doebbeling2009). This could be a part contributor to the excess CVD mortality in people with mental illness. Despite this, there is no suggestion that these interventions are less effective in people with mental illness. Although some have argued that mental illness is a valid contra-indication for surgical procedures because of compliance with follow-up care (Garg & Garg, Reference Garg and Garg2011), this view remains controversial (Lawrence & Kisely, Reference Lawrence and Kisely2010). However, there is no suggestion that mental illness is a contra-indication for prescribing pharmaceutical interventions. The gap in prescribing likely reflects inequitable access to health care screening and health service delivery. This suggests that even without new preventions or treatments for mental illness it should be possible to make gains in improving health outcomes for people with mental illness.
A downside of the GBD approach of attributing disease burden to individual disorders and risk factors is to downplay the role of comorbidity and complexity in population-level disease burden. As medical care has become increasingly specialised, greater effort is required to provide integrated and holistic health care for people with multiple health problems. Our natural desire to reduce large data sets to simple indicators mimics our tendency to seek simple solutions, even to complex problems. The literature on excess mortality in mental illness has identified a substantial range of contributing factors, from the side effects of psychotropic medications, associations between mental illness and risk behaviours such as smoking, poor diet and lack of exercise, the cognitive impacts of mental illness in influencing access to health care, the role of stigma in reducing access to health care, and silos of mental and physical health care delivery that all play a role (Lawrence & Kisely, Reference Lawrence and Kisely2010).
For instance, CVD is one of the major causes of excess death in people with mental illness. Although the GBD data could be used to ask the question of what health specialty to devote more resources to: CVD or mental health, the comparative risk assessment presented by Charlson et al. makes clear that people with mental illness experience a higher degree of burden due to CVD and vice versa a substantial proportion of the burden due to CVD is experienced by people with mental illness. Whether there is a causal association between mental illness and CVD or not, it seems clear at least from a treatment perspective and arguably from a prevention perspective as well, integrated care and treatment solutions will be an important component of improving health outcomes for both groups.
One clearly established risk factor for CVD is smoking. While it is unclear if the association between smoking and mental illness is causal and in which direction the causality might be, there is ample evidence that at least in high-income countries (HICs) the smoking rate among people with mental illness is high, and reciprocally a high proportion of those who smoke have mental illness (Lasser et al.
Reference Lasser, Boyd, Woolhandler, Himmelstein, McCormick and Bor2000). One key aspect of CVD prevention has been reducing tobacco consumption. Evidence to date suggests that tobacco control efforts in HICs have had their least success in people with mental illness (Prochaska, Reference Prochaska2011). Similar associations exist for a number of other health risk factors, including alcohol and other drugs. Whether we have evidence to determine if there is a causal link between mental illness and these health risk factors does not diminish the need to address both the health risk factor and mental illness in effective interventions.
There is established evidence that people with mental illness have high rates of comorbidities with physical health problems and health risk factors. As demonstrated by Charlson et al., the burden of excess mortality around the globe is high. There have been some early steps to address the underlying inequalities. There are a growing range of clinical guidelines and position statements (e.g., Marder et al.
Reference Marder, Essock, Miller, Buchanan, Casey, Davis, Kane, Lieberman, Schooler, Covell, Stroup, Weissman, Wirshing, HAll, Pogach, Pi-Sunyer, Bigger, Friedman, Kleinberg, Yevich, Davis and Shon2004; de Hert et al.
Reference de Hert, Dekker, Wood, Kahl, Holt and Möller2009). There have also been some small-scale trials of potentially effective interventions with promising results such as the use of peer supporters (Druss et al.
Reference Druss, von Esenwein, Compton, Rask, Zhao and Parker2010a
), and trials of integrated care models (Druss & von Esenwein, Reference Druss and von Esenwein2006; Druss et al.
Reference Druss, Zhao, von Esenwein, Bona, Fricks, Jenkins-Tucker, Sterling, Diclemente and Lorig2010b
). What is needed now is to use the GBD results and associated information, as developed by Charslon et al., to change public health priorities, to set goals for progress and to use ongoing indicator series as measures against these goals, and to implement some of the promising programs and models within an evaluation framework that improves our knowledge of effective interventions going forward.
You have
Access
The Global Burden of Disease study (GBD), now in its second iteration, has been and will continue to be very influential in shaping health priorities. By providing empirical estimates of disease burden based on robust methodologies the GBD has changed government thinking about priorities for health service delivery, and provides an empirical basis for setting priorities for service reform, funding allocations and research priorities based on evidence of burden and potential impact.
However, as highlighted by Charlson et al. (Reference Charlson, Baxter, Dua, Degenhardt, Whiteford and Vos2015) the GBD presentation does not always tell the full story. In any accounting exercise, decisions need to be made about classifications and the choice of how to summarise a wealth of data into the most important headline indicators. The GBD has focused on years of life lost (YLLs) and years lived with disability (YLDs; Whiteford et al. (Reference Whiteford, Degenhardt, Rehm, Baxter, Ferrari, Erskine, Charlson, Norman, Flaxman, Johns, Burstein, Murray and Vos2013). Charlson et al. (Reference Charlson, Baxter, Dua, Degenhardt, Whiteford and Vos2015) provide illuminating additional information about the burden of mortality due to mental illness than could be shown in the YLL summary measures presented in the main GBD reports. These data show that although YLLs directly attributed to mental illness are a cause of concern, they are small in comparison to the excess mortality observed in people with mental illness using natural history models. This gap could be even larger than Charlson et al. have calculated. Schizophrenia is the only mental disorder currently included in the cause-specific calculation of YLLs. The natural history models additionally examine excess mortality in major depressive disorder, bipolar disorder and substance use disorders, but anxiety disorders have not been included. Our own data from Western Australia (WA) found that anxiety disorders have high prevalence and at least some anxiety disorders are associated with similar levels of excess mortality as depression (Lawrence et al. Reference Lawrence, Hancock and Kisely2013). More epidemiological research is needed on the health outcomes of people with anxiety disorders before they could be included in this analysis, but if the WA results are confirmed in other places the burden of excess mortality associated with mental illness could be even higher than as stated by Charlson et al.
The high level of excess mortality associated with mental illness is one indicator of disturbing inequalities in health outcomes for people with mental illness that deserve serious attention (Thornicroft, Reference Thornicroft2013). The gap between the YLL analysis and the results of the natural history models is instructive in what it implies about the nature of the gap in life expectancy for people with mental illness and the possibility that it could be reduced even without a major new breakthrough in treating or preventing mental illness. With the most notable exception of dementia, which remains an incurable, terminal illness, there are no known biological mechanisms to believe that mental illnesses are direct causes of death. This gives hope that the excess mortality in people with mental illness is an inequality that is potentially remediable even while we do not yet know how to prevent or cure all mental illness. For instance, there are a range of evidence-based treatments for cardiovascular disease (CVD) including procedural interventions such as angioplasty, stenting and bypass grafts and pharmaceutical interventions such as aspirin, angiotensin-converting enzyme inhibitors and beta-blockers. However, people with both mental illness and CVD are less likely to receive these interventions than people with CVD who do not have mental illness (Druss et al. Reference Druss, Bradford, Rosenheck, Radford and Krumholz2000; Mitchell et al. Reference Mitchell, Malone and Doebbeling2009). This could be a part contributor to the excess CVD mortality in people with mental illness. Despite this, there is no suggestion that these interventions are less effective in people with mental illness. Although some have argued that mental illness is a valid contra-indication for surgical procedures because of compliance with follow-up care (Garg & Garg, Reference Garg and Garg2011), this view remains controversial (Lawrence & Kisely, Reference Lawrence and Kisely2010). However, there is no suggestion that mental illness is a contra-indication for prescribing pharmaceutical interventions. The gap in prescribing likely reflects inequitable access to health care screening and health service delivery. This suggests that even without new preventions or treatments for mental illness it should be possible to make gains in improving health outcomes for people with mental illness.
A downside of the GBD approach of attributing disease burden to individual disorders and risk factors is to downplay the role of comorbidity and complexity in population-level disease burden. As medical care has become increasingly specialised, greater effort is required to provide integrated and holistic health care for people with multiple health problems. Our natural desire to reduce large data sets to simple indicators mimics our tendency to seek simple solutions, even to complex problems. The literature on excess mortality in mental illness has identified a substantial range of contributing factors, from the side effects of psychotropic medications, associations between mental illness and risk behaviours such as smoking, poor diet and lack of exercise, the cognitive impacts of mental illness in influencing access to health care, the role of stigma in reducing access to health care, and silos of mental and physical health care delivery that all play a role (Lawrence & Kisely, Reference Lawrence and Kisely2010).
For instance, CVD is one of the major causes of excess death in people with mental illness. Although the GBD data could be used to ask the question of what health specialty to devote more resources to: CVD or mental health, the comparative risk assessment presented by Charlson et al. makes clear that people with mental illness experience a higher degree of burden due to CVD and vice versa a substantial proportion of the burden due to CVD is experienced by people with mental illness. Whether there is a causal association between mental illness and CVD or not, it seems clear at least from a treatment perspective and arguably from a prevention perspective as well, integrated care and treatment solutions will be an important component of improving health outcomes for both groups.
One clearly established risk factor for CVD is smoking. While it is unclear if the association between smoking and mental illness is causal and in which direction the causality might be, there is ample evidence that at least in high-income countries (HICs) the smoking rate among people with mental illness is high, and reciprocally a high proportion of those who smoke have mental illness (Lasser et al. Reference Lasser, Boyd, Woolhandler, Himmelstein, McCormick and Bor2000). One key aspect of CVD prevention has been reducing tobacco consumption. Evidence to date suggests that tobacco control efforts in HICs have had their least success in people with mental illness (Prochaska, Reference Prochaska2011). Similar associations exist for a number of other health risk factors, including alcohol and other drugs. Whether we have evidence to determine if there is a causal link between mental illness and these health risk factors does not diminish the need to address both the health risk factor and mental illness in effective interventions.
There is established evidence that people with mental illness have high rates of comorbidities with physical health problems and health risk factors. As demonstrated by Charlson et al., the burden of excess mortality around the globe is high. There have been some early steps to address the underlying inequalities. There are a growing range of clinical guidelines and position statements (e.g., Marder et al. Reference Marder, Essock, Miller, Buchanan, Casey, Davis, Kane, Lieberman, Schooler, Covell, Stroup, Weissman, Wirshing, HAll, Pogach, Pi-Sunyer, Bigger, Friedman, Kleinberg, Yevich, Davis and Shon2004; de Hert et al. Reference de Hert, Dekker, Wood, Kahl, Holt and Möller2009). There have also been some small-scale trials of potentially effective interventions with promising results such as the use of peer supporters (Druss et al. Reference Druss, von Esenwein, Compton, Rask, Zhao and Parker2010a ), and trials of integrated care models (Druss & von Esenwein, Reference Druss and von Esenwein2006; Druss et al. Reference Druss, Zhao, von Esenwein, Bona, Fricks, Jenkins-Tucker, Sterling, Diclemente and Lorig2010b ). What is needed now is to use the GBD results and associated information, as developed by Charslon et al., to change public health priorities, to set goals for progress and to use ongoing indicator series as measures against these goals, and to implement some of the promising programs and models within an evaluation framework that improves our knowledge of effective interventions going forward.