The economic burden of depression disease for China, India and the USA in 2025–2050: a health-augmented macroeconomic modelling study Authors: Meiyu Wu, Andong Li, Kehui Meng, Yangke Yi, Zixuan Zhang, Chaohao Shi, Yongtao He, Xuan Ouyang, Chongqing Tan, Xiaomin Wan Published: 2026 ABSTRACT BackgroundDepression is an increasingly severe public health challenge worldwide, causing substantial health losses and profoundly impacting national economies. As China, India and the USA bear the largest share of the global depression burden and have different developmental and demographic profiles, quantifying the condition’s long-term macroeconomic effects in these countries is crucial for gaining insight into the relationship between mental health and economic development worldwide.MethodsWe used a health-augmented macroeconomic model to assess the macroeconomic impact of depression, comparing the difference in gross domestic product (GDP) between a status quo scenario and a counterfactual scenario in which depression was assumed to be completely eliminated between 2025 and 2050. The model incorporated human and physical capital channels, capturing productivity losses, reduced labour supply and declines in household savings caused by medical expenditures. Data were drawn from multiple publicly available sources, including literature, the Global Burden of Disease Study and the World Bank.ResultsFrom 2025 to 2050, the cumulative macroeconomic burden of depression is projected to reach INT $3,340 billion across the three countries. The USA is expected to bear the largest total and per capita losses, at INT $1,416 billion and INT $4,110 per person respectively. India would follow with INT $1,100 billion and INT $762 per person, while China would account for INT $873 billion and INT $615 per person.ConclusionsAcross different economic models, depression has a significant and sustained negative impact on long-term macroeconomic development. Policymakers should take prompt action to implement effective prevention and management strategies to curb the growing health and economic impacts of depression.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12916-026-04982-x. Background Depression is a common and severe mental disorder and represents one of the leading contributors to global disability and premature mortality [1–3]. It is estimated that more than 330 million people worldwide are living with depression [4, 5]. The burden of depression is unevenly distributed across countries and regions. In absolute terms, China, India, and the USA account for the largest numbers of individuals affected by depression, together representing a substantial portion of the burden of depression worldwide [6, 7] (Additional file 1: Fig. S1 and Table S1). Beyond its high prevalence, depression is strongly associated with an increased risk of suicide and other chronic conditions, including cardiovascular disease [8, 9]. It is also linked to a reduction in life expectancy of approximately 7 years for women and 10 years for men [8, 10, 11]. In recent decades, the prevalence of depression has continued to rise, driven by factors such as increasing societal stress, changing lifestyles, and the pervasive impact of digital life, making depression a growing global public health challenge [12]. Beyond causing a significant health burden, depression imposes a substantial economic burden on societies [13, 14]. Productivity losses due to absenteeism, presenteeism, and premature exit from the labour force, together with increased healthcare utilization, place considerable strain on families, employers, and healthcare systems. Depression also generates additional economic burden through its effects on caregivers and social support networks. Despite this broad impact, depression has received insufficient attention from governments and policy makers, partly because of the limited availability of comprehensive and comparable estimates of its economic burden [12]. This issue is particularly prevalent in low- and middle-income countries (LMICs) like India and in rapidly transitioning economies such as China, where detection and treatment rates remain persistently low [15, 16]. Even in a high-income country like the United States, fewer than half of patients with depression receive adequate treatment [17, 18]. Therefore, comprehensively assessing the economic burden of depression can help policymakers develop evidence-based strategies to reduce the disease burden associated with depression. Existing studies that estimate the economic burden of depression have largely relied on cost-of-illness approach, which aggregate direct medical costs and indirect productivity losses [19–22]. Although informative, this approach is inherently static. It does not account for demographic dynamics arising from depression-related morbidity and mortality across different population subgroups, nor does It capture macroeconomic adjustment mechanisms, such as changes in savings, investment, and capital accumulation induced by increased healthcare expenditures [23]. As a result, this approach is unable to reveal the dynamic and long-term economic effects of depression on national economies. To address these limitations, our study aimed to evaluate the macroeconomic burden of depression in China, India, and the USA from 2025 to 2050 using a health-augmented macroeconomic model, incorporating demographic dynamics and changes in savings and investment behaviors due to healthcare expenditures. Methods Model description In this economic evaluation, we quantified the macroeconomic burden associated with depression in China, India, and the USA using a health-augmented macroeconomic model, which has been widely adopted in previous studies [23–27]. The model includes health-related variables such as productivity losses and medical costs, which affect macroeconomic outcomes through multiple pathways. First, depression diminishes the effective labour supply at various educational levels by the incidence of illness and mortality events. Depression reduces productivity and increases absenteeism, while depression-related mortality further decreases the working-age population. Second, families with depression patients need to spend their savings to cover out-of-pocket medical expenses. Alternatively, paying for treatment through commercial insurance would increase the burden of premiums, while paying through the national health insurance system would represent an increase in government tax for public health. Regardless of the payment source, the related medical expenditures diminish the overall savings capacity of households or society and reduce available funds for investment, thereby hindering the physical capital accumulation within the broader economy. The framework for the health-augmented macroeconomic model is shown in Fig. 1. Fig. 1Flow chart of the health-augmented macroeconomic model. Symbols in parentheses denote model parameters. Detailed definitions are provided in the supplementary materia Flow chart of the health-augmented macroeconomic model. Symbols in parentheses denote model parameters. Detailed definitions are provided in the supplementary materia The model was constructed based on a production function that incorporates two key production factors: physical capital and human capital. In the evaluation of disease economic burden, physical capital accumulation is associated with disease-related healthcare expenditures. Human capital is defined as the total effective labour supply within a specific age cohort, which is influenced by population size, labour force participation rates, educational attainment, and years of work experience. We projected two scenarios for the period 2025–2050 to assess the macroeconomic burden attributable to depression. (1) Status quo scenario: where the epidemic of depression follows current trends without any intervention; (2) Counterfactual scenario: assuming depression had been eliminated from the outset of the time frame. The counterfactual scenario calculated the accumulation of physical capital by incorporating the proportion of savings that would otherwise have been allocated to investment, but which were instead diverted towards healthcare spending. Population size and labour force participation rates would change with the elimination of diseases. The macroeconomic burden of depression was calculated as the cumulative difference in aggregate output between the two scenarios. More specific details of the model were described in the Additional file 1 [28–31]. Model parameters and data sources Data on physical capital stock (in constant 2017 international dollars), depreciation rate and the output elasticity of physical capital for each country were sourced from the Penn World Table (version 11.0) [32]. The output elasticity of physical capital measures the percentage change in output resulting from a 1% change in physical capital stock. We sourced national saving rates for each country from 2015 to 2024 from the World Bank [33]. For the projection from 2025 to 2050, we assumed that these rates would remain at their average for the period from 2015 to 2024. Per capita gross domestic product (GDP) was obtained from the World Bank, and projected growth rates for 2025–2030 were taken from the International Monetary Fund [33, 34]. Beyond 2030, GDP growth was assumed to equal the five-year average of projected rates. Treatment costs were derived from country-specific sources: a nationally representative household survey for China, a tertiary hospital-based cost-of-illness study for India, and national-level estimates based on survey and administrative data for the USA [22, 35, 36]. The fraction of treatment costs redirected to savings was assumed to equal the national saving rate, We assume that individuals have labour force participation from age 15 until retirement age. Information on country-specific population sizes and age distribution by five-year age groups were taken from the United Nations Population Division [37]. Age-specific labour force participation rates relied on the International Labour Organization database [38]. The Mincer elasticities for education and experience were obtained from published studies. Data on incidence, prevalence, years of life lost (YLL), years lived with disability (YLD), and all-cause mortality were all estimated from the global burden of disease study (GBD) database [39]. The hazard ratio for mortality among individuals with depression was obtained from published studies [40–43]. We obtained age-specific educational attainment data from the Barro-Lee Educational Attainment Database [44, 45]. Further details regarding the model parameters were described in the Additional file 1: Table S2 [46, 47]. Outcomes We reported the macroeconomic burden of depression of each country for the time period 2025 to 2050, which was calculated as the cumulative annual GDP difference between the status quo scenario and the counterfactual scenario. In order to enable comparisons across different countries, we evaluated various measures of economic performance, including per capita loss, the percentage of 2024 GDP, and the percentage of total GDP during the period from 2024 to 2050. We analysed the age distribution of the macroeconomic burden across three groups: young (15–34 years old), middle-aged (35–54 years old), and pre-retired (≥ 55 years old). All economic data and estimates were discounted at 3% per year, and were converted to 2024 International Dollars (INT $). Uncertainty analysis We conducted a series of sensitivity analyses to explore the uncertainty and assumptions of the model. First, we performed a scenario analysis in which the proportion of decreased prevalence was gradually increased from 0% to 100%, to assess the GDP benefit associated with each % prevalence averted. Furthermore, we conducted a sensitivity analysis on the discount rate, given that its appropriate value is contingent on the prevailing economic environment and varies by country. Accordingly, we adopted a 3% rate for our primary analysis and tested alternative scenarios of 0%, 5%, and 8% to assess the variance in the estimated economic burden. Finally, We used 300 Monte Carlo simulations to assess the combined uncertainty of all model parameters. Values were sampled from uniform distributions defined by a ± 50% variation around the baseline estimates. The results are presented as a 95% uncertainty interval (UI) based on the 300 iterations. Model construction and data analysis were both performed in R software (R version 4.0.5; http://www.r-project.org). Results The macroeconomic burden attributable to depression from 2025 to 2050 totaled INT $3,340 billion, with Table 1 detailing the burden across China, India, and the USA over this period. Among the three countries, the USA had the largest total cumulative economic loss at INT $1,416 billion, followed by India at INT $1,100 billion and China at INT $873 billion. In terms of per capita economic loss, the USA ranked highest at INT $4,110, while China and India had per capita losses of INT $615 and INT $762, respectively. The share of depression-related loss in 2024 GDP was highest in India at 7.73%, followed by USA at 5.51% and China at 2.60%, and for the share in 2025–2050 cumulative GDP, the USA stood at 0.156%, India at 0.113%, and China at 0.066%. Table 1Estimates of the macroeconomic burden attributable to depression for the period 2025–2050CountryEconomic loss, billions of 2024 INT$Per capita loss in 2024 INT$Percentage of 2024 GDPPercentage of GDP in 2025–2050China873 (846–1480)615 (596–1042)2.6% (2.52%−4.41%)0.066% (0.064%−0.112%)India1100 (1079–1618)762 (747–1120)7.73% (7.58%−11.36%)0.113% (0.111%−0.166%)USA1416 (1395–1888)4110 (4050–5480)5.51% (5.43%−7.35%)0.156% (0.154%−0.208%) Table 1Estimates of the macroeconomic burden attributable to depression for the period 2025–2050CountryEconomic loss, billions of 2024 INT$Per capita loss in 2024 INT$Percentage of 2024 GDPPercentage of GDP in 2025–2050China873 (846–1480)615 (596–1042)2.6% (2.52%−4.41%)0.066% (0.064%−0.112%)India1100 (1079–1618)762 (747–1120)7.73% (7.58%−11.36%)0.113% (0.111%−0.166%)USA1416 (1395–1888)4110 (4050–5480)5.51% (5.43%−7.35%)0.156% (0.154%−0.208%) Estimates of the macroeconomic burden attributable to depression for the period 2025–2050 In terms of age distribution, the macroeconomic burden of depression is concentrated in the middle‑aged group. In China and India, this group accounts for 61% and 63% of the total loss, respectively. In the USA, the middle‑aged group contributes 50% of the loss, followed closely by the pre‑retired group with 43%. The young group accounts for the smallest share in all three countries (China 13%, India 10%, USA 7%). The age distribution results are presented in Fig. 2. Fig. 2Age distribution of the macroeconomic burden of depression in China, India, and the USA (2025–2050). Shares of total GDP loss by three broad age groups: young (15–34 years old), middle-aged (35–54 years old), and pre-retired (≥ 55 years old) Age distribution of the macroeconomic burden of depression in China, India, and the USA (2025–2050). Shares of total GDP loss by three broad age groups: young (15–34 years old), middle-aged (35–54 years old), and pre-retired (≥ 55 years old) Evaluation results showed that treatment costs accounted for a significantly higher proportion of the macroeconomic burden associated with depression in the USA, reaching 46%. However, this proportion was substantially lower in China at 19% and even lower in India at 9%. The contribution of treatment costs to the total estimate is presented in Fig. 3. Fig. 3Percent contribution of treatment costs to total estimate Percent contribution of treatment costs to total estimate We adjusted the percentage reduction in depression prevalence in order to assess the GDP benefits per percentage point reduction. The results indicate that, as prevalence declines, the GDP benefits across the three economies exhibit an approximately linear growth trend (Fig. 4). A 10% reduction in prevalence averted INT $87 billion, INT $110 billion and INT $142 billion respectively; at a 50% reduction in prevalence, the averted GDP losses increased to INT $437 billion, INT $550 billion and INT $708 billion respectively. Across all intervention intensities, the USA demonstrated the largest absolute GDP gains, followed by China and India. Fig. 4GDP benefit associated with averted disease prevalence. GDP benefit associated with averted disease prevalence. The total macroeconomic burden of depression on the three countries between 2020 and 2050 changes significantly with variations in the discount rate, as shown in Table 2. The economic burden for China decreased from INT $1,328 billion at a 0% discount rate to INT $681 billion at a 5% discount rate and INT $489 billion at an 8% discount rate. India’s burden decreased from INT $1,871 billion to INT $793 billion and INT $505 billion; and the USA’s burden dropped from INT $2,280 billion to INT $1,060 billion and INT $715 billion. Notably, the USA had the largest economic burden at each discount rate, followed by India and China, indicating that discount rate choice substantially impacts long-term macroeconomic burden estimates of depression, with higher discount rates reducing the present value of future economic losses. More scenario analysis results were presented in Additional file 1: Figs. S2, S3 and S4. Table 2Total macroeconomic burden attributable to depression from 2025 to 2050 using different discount ratesCountryEconomic burden discounted at 0% in billions of 2024 INT$Economic burden discounted at 5% in billions of 2024 INT$Economic burden discounted at 8% in billions of 2024 INT$China1328 (1297–2244)681 (665–1149)489 (478–827)India1871 (1825–2732)793 (783–1164)505 (501–744)USA2280 (2377–3013)1060 (1149–1413)715 (759–961) Table 2Total macroeconomic burden attributable to depression from 2025 to 2050 using different discount ratesCountryEconomic burden discounted at 0% in billions of 2024 INT$Economic burden discounted at 5% in billions of 2024 INT$Economic burden discounted at 8% in billions of 2024 INT$China1328 (1297–2244)681 (665–1149)489 (478–827)India1871 (1825–2732)793 (783–1164)505 (501–744)USA2280 (2377–3013)1060 (1149–1413)715 (759–961) Total macroeconomic burden attributable to depression from 2025 to 2050 using different discount rates Discussion This study provides a dynamic, economy-wide assessment of the macroeconomic burden of depression in China, India, and the USA between 2025 and 2050. Using a health-augmented macroeconomic model, our findings indicate that depression is projected to generate cumulative output losses exceeding INT $3.40 trillion over the next 26 years across the three countries. The USA is expected to have the largest absolute loss, at INT $1,416 billion, followed by India with INT $1,100 billion and China with INT $873 billion. Although China and India have a larger number of people living with depression, the model indicates that the USA will incur the largest economic losses. This finding underscores that the economic burden of depression is not determined by prevalence alone, but by how illness interacts with high-productivity labour and capital-intensive economies. In the USA, the contribution of treatment costs to the total macroeconomic loss accounts for nearly half, indicating that high depression-related healthcare expenditures have significantly crowded out savings and investment, consequently constraining the accumulation of physical capital [48, 49]. Moreover, productivity losses among a highly educated workforce amplify the economic consequences of morbidity and premature mortality, leading to disproportionate reductions in effective human capital [50, 51]. In contrast, depression exerts its macroeconomic impact in China and India primarily through labour supply and demographic channels. Long-term growth in these economies has relied heavily on demographic dividends [52, 53], which are increasingly undermined by the high prevalence of depression among working-age populations and persistently low treatment rates [54, 55]. In China, where high household and national savings have been central to sustaining investment-led growth [56], depression-related medical expenditures and productivity losses further erode saving capacity, weakening a key pillar of long-term economic expansion. In India, the relatively small contribution of treatment costs reflects limited access to care rather than a lower disease burden, implying that much of the economic loss arises from untreated morbidity and premature mortality rather than healthcare spending. Distance to facilities, waiting times, and out-of-pocket costs have been identified as major barriers to public healthcare utilization in India [57]. Moreover, estimates indicate substantial inefficiency in India’s public healthcare system, with approximately 40% of resources not being used effectively [58]. Although India had the highest share of depression-related loss in 2024 GDP, its share in cumulative GDP over 2025–2050 was lower than that of the USA. This is mainly because India’s faster projected GDP growth enlarges the cumulative denominator, lowering its long-term share. These findings highlight an important conceptual distinction between treated and untreated disease burdens in macroeconomic analysis. In high-income settings, depression generates economic losses largely through the financial and productivity costs of treatment and managed illness. In emerging economies that depend on demographic advantages and high savings, the burden is dominated by foregone labour supply and human capital losses resulting from unmet need for care. In the context of slowing global growth and population ageing, depression exerts a persistent drag on economic development by systematically reducing labour force participation, eroding effective human capital, and weakening savings and investment capacity. It is evident that the long-term productivity losses associated with depression are increasingly eroding the growth gains that would otherwise be generated by technological progress and capital accumulation. Notably, the per capita economic loss in the USA is substantially higher than in China and India. This difference is likely driven by the higher income level and higher treatment costs per patient in the USA. Part of these high treatment costs may reflect healthcare system inefficiencies, such as higher drug prices and administrative overhead, in addition to the intrinsic severity of depression. This suggests that cross-country differences in economic burden may partly arise from differences in healthcare system efficiency, not solely from differences in disease prevalence or severity. The concentration of the economic burden in the middle-aged group is attributable to this cohort’s highest labour force participation, greatest human capital, and highest average earnings. Accordingly, depression-induced labour supply or productivity losses generate the largest GDP losses in this age segment. The markedly higher burden share of the pre-retired group in the USA relative to China and India reflects the higher labour force participation and later retirement age among older Americans. These findings highlight the need for age-targeted mental health policies. Our results indicate an approximately linear relationship between reductions in depression prevalence and macroeconomic outcomes, with no clear indication of diminishing returns over the range considered. This pattern is consistent with the model’s first-order approximation of labour supply effects under low mortality [24]. Even moderate interventions, such as expanding mental health screening coverage, improving access to psychotherapy, and increasing medication coverage, could yield sustained gains in productivity and economic output, with stable marginal benefits for each percentage reduction in prevalence. For countries with limited resources that are unable to implement large-scale interventions, targeting high-risk groups in order to rapidly reduce prevalence may be a more cost-effective strategy. More broadly, the positive relationship between health outcomes and economic growth has been documented in various regions, including South Asia [59]. To our knowledge, this study is the first to quantitatively assess the macroeconomic burden of depression. The model used captures reductions in population size and labour force participation as responses to changes in mortality and morbidity, treats education and work experience as key determinants of average human capital, and considers how treatment-related expenditures affect household saving and investment behavior. Our study quantifies the impact of depression on economic development in China, India, and the USA over the next 26 years, providing decision-making references for mental health policy formulation and macroeconomic planning in these three countries. Beyond national-level implications, our findings are also relevant to global economic stability and inequality. Together, China, India and the USA account for a significant proportion of global GDP and depression-related DALYs. Persistent depression-related productivity losses in these economies could therefore impact global supply chains, labour markets and capital flows. Furthermore, the differential capacity to absorb and mitigate mental health burdens could exacerbate economic inequalities both between and within countries, particularly given that mental disorders disproportionately affect working-age populations. The objective of this study is to quantify the economic loss imposed by depression on the overall economy, thereby providing policymakers and the public with a clearer understanding of the societal harm caused by depression from an economic perspective. Our estimates should be interpreted as macroeconomic burden estimates attributable to depression, based on a counterfactual scenario in which depression is absent. As in previous applications of the Bloom framework, intervention costs required to achieve such a scenario were not explicitly modelled. Therefore, our results do not represent net economic returns of specific treatment or prevention programmes, but rather the potential macroeconomic losses associated with depression under the current burden trajectory. Following previous studies, we assume that the proportion of freed-up treatment costs converted into savings equals the national saving rate. However, health financing systems vary across countries, which may affect the transmission from treatment costs to savings. This simplification should be considered when interpreting cross-country comparisons of the estimated economic burden. In addition, unlike diseases such as COPD, cancer, and cardiovascular disease, for which YLL are directly available from GBD, our model adopts an attribution method to estimate YLL associated with depression. This approach has been used in macroeconomic burden studies of air pollution-related diseases, tobacco-related diseases, and Alzheimer’s disease to estimate risk factor-attributable YLL [60–62]. There are some limitations in our study. First, the model relies on several key assumptions, such as the independence of mortality and morbidity across diseases, the constancy of saving behavior. Although these assumptions are common in health-augmented macroeconomic models [23–27], they may not fully capture complex real-world interactions. Second, we did not account for informal caregiving provided by family members, which could lead to additional labour supply losses and thus an underestimation of the true economic burden. Third, while we incorporated changes in savings due to healthcare expenditures, the model does not endogenize household or government behavioral responses to health shocks, such as adjustments in education investment or public health financing, which could influence long-term economic outcomes. In addition, our model treats depression-related healthcare spending as a reduction in savings and investment and does not capture potential feedback effects (e.g., reinvestment of healthcare-related revenues). As a result, the estimated economic burden reflects the reallocation and crowding-out effects associated with depression and may not fully capture broader macroeconomic feedback mechanisms. Fourth, data limitations for certain parameters (e.g., country-specific treatment costs and labour participation rates by education and health status) necessitated the use of extrapolations or assumptions based on international averages, which may not fully reflect national contexts. For India, depression treatment costs were derived from a single tertiary academic hospital due to the lack of nationally representative data, which may limit the generalizability of our cost estimates for the Indian population. Fifth, our model does not account for the spillover effects of mental health interventions on other diseases, such as the reduced risk of cardiovascular diseases among patients with depression, which could further decrease healthcare costs and enhance productivity. Sixth, depression is characterised by episodic, relapsing-remitting patterns. Our model simplifies the probability that a patient has not yet recovered from depression at time t, which may not fully capture recurrence, remission, and heterogeneous episode duration. Although the use of GBD prevalence data partially reflects active disease states at the population level, some uncertainty remains regarding the cumulative productivity effects of recurrent depression. Future research could apply more refined approaches, such as Markov switching models, to better represent the natural history of depression. Finally, our model follows the framework proposed by Bloom et al., which treats disease prevalence as exogenous and models only the one-way effect of disease on GDP through labour supply, productivity, and capital accumulation. This approach has often been regarded as yielding conservative estimates of disease burden. However, for depression, the relationship with economic development may be bidirectional: while depression can reduce GDP, economic growth may also affect depression prevalence through offsetting channels such as improved living conditions and healthcare access versus greater stress, inequality, and social isolation. Owing to the lack of robust and generalisable estimates of this feedback effect, we did not incorporate it into the model. Therefore, our results should not be interpreted as strict lower bounds of the macroeconomic burden of depression, but rather as conservative estimates under a partial-equilibrium framework. Furthermore, our model treated technological level as exogenous owing to data limitations and therefore did not capture the possibility that depression may also reduce economic output through presenteeism, impaired work quality, or other productivity-related channels. Conclusions In conclusion, by using a health-augmented macroeconomic model, our findings indicate that under the status quo, the macroeconomic burden of depression in China, India, and the USA is projected to be substantial over the coming decades. Policymakers should take immediate action to curb the escalating health and economic impacts of depression. Investing in strategies to prevent and manage depression is not merely a financial cost, but also sustains productivity, supports capital formation resulting in substantial economic benefits. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1: Additional file 1. More details of the model construction and input parameters. Table S1. Key background indicators for depression in China, India, and the USA. Table S2. Parameter values and data sources. Figure S1. Prevalence of depression in 2021. Figure S2. The macroeconomic effects of depression under different economic growth rates in India. Figure S3. The macroeconomic effects of depression under different economic growth rates in China. Figure S4. Result of one way sensitivity analysis. Supplementary Material 1: Additional file 1. More details of the model construction and input parameters. Table S1. Key background indicators for depression in China, India, and the USA. Table S2. Parameter values and data sources. Figure S1. Prevalence of depression in 2021. Figure S2. The macroeconomic effects of depression under different economic growth rates in India. Figure S3. The macroeconomic effects of depression under different economic growth rates in China. Figure S4. Result of one way sensitivity analysis.