The global burden of anxiety disorders: challenges in calculating prevalence and burden of disease using anxiety disorders as a case-study

Abstract

Background Effective decision making in health care relies upon evidence-based measures of relative disease frequency, mortality and morbidity. Findings from the first Global Burden of Disease (GBD) study showed for the first time the high level of health loss, or ‘burden’, associated with mental disorders relative to other high priority public health issues. A major contribution to the mental health burden comes from anxiety disorders yet there has been no comprehensive assessment of the global frequency or extent of health loss from this group of common mental disorders. Accurate estimates for the prevalence and burden of anxiety disorders, relative to other health issues, across populations and over time, are needed to inform policy decisions around resource re-allocation, service provision and research priorities. The new Global Burden of Disease Study (GBD 2010) aimed to produce estimates of fatal and non-fatal disease burden using improved data and modelling techniques for a more complete range of disease and injury categories. This new initiative provided an opportunity to produce, for the first time, estimates for the global prevalence and burden of anxiety disorders. This thesis addresses the data input and methods required to synthesise populationrepresentative measures of health loss to inform health policy, and the application of these in calculating the prevalence and burden of anxiety disorders. Methods A series of systematic reviews were conducted for epidemiological studies reporting prevalence, incidence, remission and excess mortality for all mental disorders captured in GBD 2010. I collated and assessed information on data availability and study quality, identifying strengths and weaknesses in the current research and calculated regional coverage for prevalence data. Then, focusing on anxiety disorders, I used random effects meta-regression to explore sources of variability (methodological and substantive) in the prevalence data. Next, a global prevalence model of anxiety disorders was derived using a Bayesian meta-regression tool (DisMod-MR). The model input included data sourced iii through the systematic review and expert input. The data were adjusted for methodological differences identified in the preliminary meta-regression and then estimates derived for populations where data were lacking based on epidemiologically similar populations and country-level covariates such as exposure to conflict. Years of life lived with disability (YLDs) were calculated based on prevalent cases multiplied by an average disability weight taken from the weights allocated to severity proportions (mild, moderate and severe). YLDs and disability-adjusted life years (DALYs) were calculated by sex and age across 21 world regions for 1990, 2005 and 2010. The proportion of burden attributable to anxiety disorders was compared with other mental and physical disorders to provide an objective evidence base to inform decision making in global health policy and service provision. Findings Anxiety disorders were common, with an average point prevalence of 4.0% (95%UI 3.8– 4.3%) and were associated with a high degree of health loss due to the disability resulting from these disorders. In 2010, they accounted for 390 YLDs/DALYs per 100,000 persons (95%UI 191–371/100,000), making anxiety disorders the sixth leading cause of global disability, in terms of YLDs. While there was no direct mortality associated with anxiety disorders in GBD 2010 7% of suicide deaths were attributed to anxiety disorders. Between 1990 and 2010, crude YLDs from anxiety disorders increased by 36%. There was no detectable change in the prevalence of anxiety disorders over the past two decades. The increase in absolute YLDs was explained by global population growth and population ageing. Females accounted for 65% of the DALYs caused by anxiety disorders. The age pattern was similar for both sexes, with YLD rates peaking in the 15–34-year age group: 387 YLDs per 100,000 males and 721 YLDs per 100,000 females. Higher YLD rates were found in high-income GBD regions (North America High Income, Australasia, and Southern Latin America) and in regions with populations exposed to conflict (North Africa/Middle East and Central Europe), while lower rates were found in East Asia. There was considerable uncertainty around estimates, particularly for regions where no empirical data were available. Implications Anxiety disorders are chronic, disabling conditions, and the findings reported here show for the first time their public health significance. To respond to the high burden and low treatment rates of anxiety disorders, policy makers and clinicians in all countries need to ensure cost-effective interventions are widely available. Estimates of prevalence and disease burden could be improved by strengthening the quality of epidemiological data, and identifying whether variations in regional distributions of anxiety disorders are substantive or an artefact of cultural or methodological differences.