In 2023, approximately six out of ten primary care patients reported ever having COVID‑19. In total, 107 011 primary care patients participated in the PaRIS survey in 2023, for which complete data were available for 103 000. Approximately 56 000 reported ever testing positive for or being diagnosed with COVID‑19. Among these, approximately 7 800 reported having symptoms persisting beyond three months, consistent with a Long COVID diagnosis (see Figure 3). Those who did not report being diagnosed with COVID‑19 were classified as not having Long COVID.
The prevalence and impact of Long COVID in the primary care population
Findings from the OECD PaRIS survey
Note: Long COVID status refers to having ever had Long COVID.
Source: OECD PaRIS 2024 Database.
Approximately 7% of primary care patients surveyed in 2023 reported ever having Long COVID. Overall, among all the primary care patients surveyed, 54.4% reported having tested positive for or being diagnosed with COVID‑19. Among those who had experienced COVID‑19, 13.9% reported persistence of symptoms of Long COVID beyond 3 months (adjusted1 estimate 13.9%, (CI 12.3‑15.7)), and 6.5% reported persistent symptoms beyond 12 months. The overall Long COVID prevalence reported among primary care patients in the PaRIS survey was 7.6% (adjusted estimate 7.2% (CI 6.3‑8.2)). Overall, in 2023, 5.6% (adjusted estimate 5.1% (CI 4.5‑5.9)) reported persisting Long COVID symptoms, and 3.5% (adjusted estimate 3.2% (CI 2.8‑3.8)) of primary care patients reported persistence of Long COVID beyond 12 months.
The proportion of primary care patients reporting ever having COVID‑19 ranged from 32.6% in Romania to 77.7% in Iceland. The proportion of COVID‑19 infected patients who reported developing Long COVID ranged from 8.0% (CI 6.8‑9.4) in Greece to 22.9% (CI 20.3‑25.8) in Italy. Long COVID prevalence estimates among primary care patients ranged from 3.9% (CI 3.3‑4.7) in Greece to 10.8% (CI 9.7‑12.0) in Norway (see Figure 4). Patients who reported persistence of Long COVID symptoms beyond 12 months ranged from 1.7% (CI 1.3‑2.1) in Greece to 4.5% (CI 3.7‑5.3) in Iceland.
Note: Adjusted estimates using the multilevel model, with comparative intervals for the estimates indicated by the span of the bars in grey. No estimate was available for the United States as American PaRIS participants were not asked the questions related to Long COVID.
Source: OECD PaRIS 2024 Database.
In PaRIS, Long COVID is a frequent complication reported by on average 14% of patients with prior COVID‑19 across OECD countries. This is a similar proportion to that reported in a cohort study of COVID‑19 patients in the Netherlands (13%) in 2021 (Ballering et al., 2022[9]), in the United States in 2021/22 (14.6%) (Perlis et al., 2023[10]), and higher than in a UK cohort study in 2021 (5.4%) (Subramanian et al., 2022[4]).
The estimated prevalence of ever having had Long COVID of 7.2% in the primary care population represents a considerable disease burden, and is highly consistent with the 6.9% estimate from a the United States survey in 2023 (Fang, Ahrnsbrak and Rekito, 2024[11]; Adjaye-Gbewonyo et al., 2023[12]). Earlier studies from 2022 in France (4.0%) (Coste et al., 2023[13]), Germany (5.4%) (Diexer et al., 2024[14]), Scotland (6.6%) (Hastie et al., 2023[15]) and the United Kingdom (5.8%) (Whitaker et al., 2022[16]) reported estimates similar to the 5.6% with persisting Long COVID symptoms in the PaRIS survey.
The patient-reported prevalence of Long COVID among OECD countries varies by a 7 percentage point difference. This likely reflects variation in the epidemiology of COVID‑19 incidence, different SARS‑CoV‑2 (the virus that causes COVID‑19) variants across waves of the pandemic, and variation in COVID‑19 vaccination uptake between countries. For example, the risk of Long COVID declined with subsequent waves of wild-type, alpha, delta and omicron variants of SARS‑CoV‑2 infections during the pandemic (Xie, Choi and Al-Aly, 2024[17]). Re‑infection with SARS‑CoV‑2 and not being vaccinated against COVID‑19 are associated with increased risk of Long COVID (Wang et al., 2025[18]).
Additionally, information bias linked to increased recognition and self-reporting of Long COVID symptoms likely also accounts for differences in reported prevalence between countries. Countries where patients are more aware of and healthcare professionals better at recognising Long COVID are more likely to self-report higher prevalence. Reporting a prior diagnosis of COVID‑19 is a necessary component of the case definition of Long COVID, therefore international differences in testing for and awareness of COVID‑19 also explain reported differences in Long COVID prevalence between countries (see Figure A.1 for correlation). The multilevel model overall prevalence estimate (7.2%) differed slightly from the crude overall prevalence estimate (7.6%). This difference may be attributed to the country level effect, the primary care provider effect, and the effect of age and sex standardisation. The observed variation in country-specific prevalences may account for most of this difference. A possible explanation is that the probability of self-reporting Long COVID symptoms is influenced by differences in the education level and awareness of Long COVID among primary care patients between different countries (see Table 1).
The persistence of symptoms beyond 12 months reported by 3.5% of primary care patients constitutes the potential long-term disease burden of the COVID‑19 pandemic, supported by estimates that only 7‑10% of Long COVID patients recover after two years of follow-up (Al-Aly et al., 2024[19]).
← 1. Estimates adjusted by age, gender and country variation using the multilevel model.