This chapter presents the estimated burden of long COVID on health systems and the economy. Drawing on results from an OECD microsimulation model, it details estimates of the past and projected prevalence of long COVID in the general population across OECD and EU Member countries, and its effects on health expenditure. The analysis also examines the wider economic implications of the disease, including its impact on labour markets.
Addressing the Costs and Care for Long COVID
The Long Shadow of the Pandemic
Abstract
3.1.1. In 2021, prevalence of long COVID was estimated at 5.3% of the total population in OECD and EU countries
The estimated prevalence of long COVID across 42 countries in 2021 is presented in Figure 3.1, showing substantial variation in how widely the condition initially affected populations. The lowest prevalence rates were observed in New Zealand (0%), Japan (0.1%) and Norway (0.1%). These very low estimates reflect limited circulation and/or more effective management of COVID‑19 cases and prevention of post-COVID conditions in these countries.
In contrast, the highest prevalence rates were recorded in Bulgaria (16.6%), Hungary (14.2%) and Czechia (11.6%). Based on their respective population sizes, this means that approximately 1.1 million people in Bulgaria, 1.4 million people in Hungary and 1.2 million people in Czechia lived with long COVID in 2021 in those countries. These figures indicate a particularly heavy burden in parts of Central and Eastern Europe.
Average prevalence rates were around 5.3% in 2021 across both EU and OECD countries. When applied to the total population across OECD countries, this translates to an estimated 75 million people affected by long COVID in 2021.
Note: Mortality data were not reported to WHO by KOR or JPN for 2021. GBR reported low numbers of deaths for 2020-2022, possibly due to the definition of COVID‑19‑related deaths used. As a result, prevalence is underestimated for these three countries.
Source: OECD SPHEP modelling.
3.1.2. Depending on the future dynamics of COVID‑19, long COVID prevalence could stabilise around 1% of the population in the coming years
Figure 3.2 shows the estimated prevalence of long COVID across OECD countries over time under three distinct scenarios (see Chapter 2, Section 2.2). Under the baseline scenario, the prevalence of long COVID peaks in 2021 and then declines steadily, reaching very low levels by 2026 and remaining close to zero for the remainder of the projection period. This scenario is considered the most conservative and, as such, is likely to underestimate the potential future burden of long COVID.
In contrast, the scenario that assumes a continued 5% incidence of COVID‑19 beyond 2023 projects a higher residual prevalence of long COVID. Here, prevalence does not fall as sharply; instead, it stabilises at around 0.6% of the OECD population, fuelled by a continuous inflow of new cases each year. The scenario with a 10% continued COVID‑19 incidence projects a starker picture, with long COVID prevalence remaining above 1% throughout the projected years.
Because the baseline scenario assumes no new cases after 2023, it provides the lowest and most optimistic (but unlikely) counter-factual scenario, suggesting that unless circulation of SARS‑CoV‑2 ceases, the burden of long COVID may persist at important levels for years into the future.
Source: OECD SPHEP modelling.
Long COVID direct healthcare costs represented a relatively modest but non-negligeable share of healthcare spending across most countries in 2021 when the prevalence peaked, with estimates ranging from near zero to approximately 1.9% of total health expenditures (Figure 3.3).
The highest burden in 2021 was observed in Bulgaria, where long COVID accounted for approximately 1.9% of total health expenditure. Hungary, Latvia and the Slovak Republic also had elevated spending levels, with long COVID representing between 1.2% and 1.6% of their total health budgets. The United States had one of the highest per capita costs at around USD 115 per capita, although this represents less than 1% of total health expenditure in 2021 due to the country’s substantially larger overall health spending base. At the lower end of the spectrum, Japan, Denmark, the United Kingdom and New Zealand showed minimal impacts, with long COVID representing less than 0.1% of total health expenditure and per capita costs below USD 5.
Both OECD and EU averages show long COVID direct healthcare costs in 2021 representing approximately 0.6‑0.8% of total health expenditure, with per capita costs around USD 22 for EU and USD 21 for OECD countries. Using the total OECD population, total healthcare costs across OECD countries in 2021 are estimated around USD 53 billion.
Note: Mortality data were not reported to WHO by KOR or JPN for 2021. GBR reported low numbers of deaths for 2020-2022. As a result, prevalence and therefore direct health expenditure is underestimated for these three countries.
Source: OECD SPHEP modelling.
The long-term direct healthcare costs were estimated under the scenarios presented in Chapter 2, Section 2.2. Figure 3.4 captures both the acute phase of the pandemic’s impact on healthcare spending and various long-term projections based on different assumptions about the persistence of long COVID cases.
All three scenarios show a similar initial pattern, with long COVID costs rising sharply to peak at around 0.63% of direct healthcare expenditure in 2021 in OECD countries, reflecting the cumulative burden of infections during the pandemic’s most intense period with particularly high COVID‑19 incidence and virulent variants of the virus. Following this peak, all scenarios show a decline through 2022-2025 as the acute phase subsides and initial cases either resolve or stabilise. However, the trajectories diverge significantly after 2024, reflecting the different assumptions made about the long-term incidence of COVID‑19.
Under the most conservative scenario (with no new long COVID cases from 2024), costs decline nearly to zero by 2027 and remain at essentially negligible levels through 2035, as long COVID cases fully resolve and new COVID‑19 infections cease to contribute meaningfully to the healthcare burden. In contrast, the two scenarios incorporating residual long COVID cases show persistent costs stabilising at different plateaus: the low residual transmission scenario maintains costs at approximately 0.07% of total health expenditure, while the most pessimistic scenario stabilises at a higher level of around 0.14% of total health expenditure, which amounts to approximately USD 11 billion in total per year across OECD countries.
Source: OECD SPHEP modelling.
Figure 3.5 illustrates the impact of long COVID on labour-force reductions across EU and OECD Member countries in 2021, expressed as a percentage change from the baseline. The estimated reduction in effective labour-force capacity includes the combined effects of sick leave, workforce exits, absenteeism and presenteeism among individuals suffering from long COVID symptoms. For each country, the confidence intervals show a range for sick leave duration of 0‑12 weeks, with a baseline of 6 weeks (see Chapter 2, Section 2.2).
The largest labour-force reductions are observed in Central and Eastern European countries (Bulgaria, Hungary and the Slovak Republic), where long COVID is estimated to have reduced effective labour capacity by approximately 2.1‑3.0% in 2021. These countries show substantial negative impacts with relatively large uncertainties, indicating both significant disease burden and measurement challenges. At the opposite end of the spectrum, countries like Japan, Norway and Australia show the smallest impacts, with labour-force reductions of approximately 0.01‑0.02%.
The OECD and EU averages both show labour-force reductions of approximately 1%. These benchmark figures indicate that across developed economies, long COVID had a measurable and economically meaningful impact on labour supply and productivity in 2021. The combination of workforce exits, sick leave and reduced productivity through absenteeism and presenteeism represents a significant macroeconomic shock, with implications for GDP growth, wage pressures and economic recovery in the post-acute phase of the pandemic.
Note: Mortality data were not reported to WHO by KOR or JPN for 2021. GBR reported low numbers of deaths for 2020-2022. As a result, prevalence and therefore labour force reduction is underestimated for these three countries.
Source: OECD SPHEP modelling.
Figure 3.6 shows how workforce disruptions from long COVID translated into macroeconomic costs in 2021, with substantial variation across nations reflecting differences in both labour-market impacts and underlying economic structures.
Mirroring the impact on the labour force, the highest GDP losses in percentage terms are observed in Bulgaria, Hungary and the Slovak Republic, where long COVID-related labour-force reductions are estimated to have cost approximately 2.1‑3.0% of GDP in 2021. However, when measured in absolute per capita terms, the picture shifts somewhat: the United States shows the highest per capita GDP loss at approximately USD 1 256 per person in 2021. This discrepancy reflects the substantially higher US baseline GDP per capita levels. At the lower end of the spectrum, countries like Japan, Norway and Australia experienced relatively modest impacts, with GDP losses below 0.02% and per capita costs under USD 15, consistent with their smaller labour-force reductions.
The OECD and EU averages show GDP losses of approximately 0.9‑1.0% in 2021, with per capita costs of around USD 323. Applied to the OECD total population, these losses translate into an economic loss of about USD 680 billion, which is 13 times higher than the estimated direct healthcare costs.
Note: Mortality data were not reported to WHO by KOR or JPN for 2021. GBR reported low numbers of deaths for 2020-2022. As a result, prevalence and therefore productivity loss is underestimated for these three countries.
Source: OECD SPHEP modelling.
Figure 3.7 illustrates how long COVID could continue to represent a substantial burden on the economies of OECD and EU Member countries, from the initial COVID‑19 pandemic years through various long-term recovery scenarios based on different assumptions about the persistence of labour-market effects. All three scenarios begin with similar initial conditions, showing GDP losses of approximately 0.9% in 2021, reflecting the acute impact of early long COVID cases on workforce capacity. All scenarios reflect the modelling assumptions of a lower, stable risk of COVID‑19 infection in future, and a fixed duration of long COVID symptoms (see Chapter 2, Section 2.2). The losses decline sharply through 2022-2023 as the most severe phase of the pandemic recedes, dropping to around 0.3‑0.4% by late 2023. However, the trajectories diverge significantly from 2024 onward, reflecting the variation in the persisting impact of long COVID on labour markets under different scenarios. The relative magnitude of OECD estimates is consistent with previous estimates on the impact on the Australian economy equivalent to 0.5% of GDP earlier in 2020-2021 (Costantino et al., 2024[1]), declining to a lower impact of 0.26% of GDP later in 2022 (Angeles et al., 2024[2]).
Under the baseline scenario (the most conservative), GDP losses decline nearly to zero by 2025-2026, and remain at essentially negligible levels (approximately 0.01%) through 2035. This suggests an assumption that long COVID cases largely resolve, affected workers return to pre‑pandemic productivity levels, and new COVID‑19 infections cease to generate significant economic drag. In contrast, the two scenarios incorporating residual long COVID effects show persistent GDP losses stabilising at different plateaus: the low residual transmission scenario maintains losses at approximately 0.1% of GDP each year, while the moderate residual transmission scenario stabilises at a higher level of around 0.2% of GDP through the 2030s. These estimates represent up to USD 135 billion in total annually across OECD countries, roughly equivalent to the annual health budget for Spain or the Netherlands. The sustained economic costs in these scenarios underscore the potential for long COVID to represent a lasting structural change to labour markets and economic productivity, rather than merely a temporary pandemic-era disruption.
Source: OECD SPHEP modelling.
Although the macroeconomic losses associated with long COVID were most pronounced in 2021, its long‑term economic impact remains significant and broadly comparable to other major chronic conditions. For instance, projections indicate that, even under realistic recovery scenarios, the level of sustained economic loss aligns closely with the national burden typically linked to multiple sclerosis (MS), estimated at 0.1‑0.5% of GDP across OECD countries – around 0.3% in Italy (EUR 4.8 billion), 0.1% in France (EUR 2.7 billion) and 0.4% in the United States (USD 85.4 billion) (Bouleau et al., 2022[3]; Bebo et al., 2022[4]; Battaglia et al., 2022[5]).
The estimated losses also approach the lower range of the economic burden of stroke, representing 0.3‑0.4% of GDP in Europe (EUR 60 billion in 2017) and 0.5% in the United States (USD 103.5 billion in 2016) (Luengo-Fernandez et al., 2020[6]; Girotra et al., 2020[7]).
However, while chronic conditions such as stroke and MS generate most of their costs through healthcare spending and informal care, the continuing burden of long COVID arises mainly from reduced labour participation and productivity losses, reflecting a broader macroeconomic impact rather than medical expenditure alone. Overall, in proportional terms, the projected 0.1‑0.2% GDP loss from long COVID places it among the most economically significant chronic conditions in OECD and EU Member countries.
Lastly, the estimates in this report probably understate the true burden of long COVID. The condition is likely to have wide‑reaching consequences that are not yet fully understood. Beyond the core symptoms of long COVID, infection with SARS‑CoV‑2 increases the risk of developing a range of chronic conditions – including cardiovascular conditions, diabetes, neurological impairments and autoimmune disorders – which will add further pressure on health systems and increase costs in the years ahead. These effects may take considerable time to become fully visible. In addition, long COVID may affect children’s development and educational attainment – factors not yet captured in current economic projections.
References
[2] Angeles, M. et al. (2024), “The economic burden of long COVID in Australia: more noise than signal?”, Medical Journal of Australia, Vol. 221/S9, https://doi.org/10.5694/mja2.52468.
[5] Battaglia, M. et al. (2022), “Patients with multiple sclerosis: a burden and cost of illness study”, Journal of Neurology, Vol. 269/9, pp. 5127-5135, https://doi.org/10.1007/s00415-022-11169-w.
[4] Bebo, B. et al. (2022), “The Economic Burden of Multiple Sclerosis in the United States: Estimate of Direct and Indirect Costs”, Neurology, Vol. 98/18, pp. E1810-E1817, https://doi.org/10.1212/WNL.0000000000200150.
[3] Bouleau, A. et al. (2022), “The socioeconomic impact of multiple sclerosis in France: Results from the PETALS study”, Multiple Sclerosis Journal - Experimental, Translational and Clinical, Vol. 8/2, https://doi.org/10.1177/20552173221093219.
[1] Costantino, V. et al. (2024), “The public health and economic burden of long COVID in Australia, 2022–24: a modelling study”, Medical Journal of Australia, Vol. 221/4, pp. 217-223, https://doi.org/10.5694/mja2.52400.
[7] Girotra, T. et al. (2020), “A contemporary and comprehensive analysis of the costs of stroke in the United States”, Journal of the Neurological Sciences, Vol. 410, https://doi.org/10.1016/j.jns.2019.116643.
[6] Luengo-Fernandez, R. et al. (2020), “Economic burden of stroke across Europe: A population-based cost analysis”, European Stroke Journal, Vol. 5/1, pp. 17-25, https://doi.org/10.1177/2396987319883160.