The Scale of the Problem
For patients: read the sections on personal economic impact (lost income, costs) to understand the financial dimension of the illness.
For caregivers: read the caregiver-burden and cost sections to understand the economic strain on families.
For clinicians: read the economic-impact sections to understand the broader burden of the disease on patients and society.
For researchers: read the Scale of the Problem and Evidence Gaps and Research Directions sections to inform health-economic research.
ME/CFS imposes economic costs that are systematically underestimated, because the disease itself is systematically underdiagnosed. An estimated 84–91% of U.S. patients remain undiagnosed (Committee on the Diagnostic Criteria for Myalgic Encephalomyelitis/Chronic Fatigue Syndrome 2015) — every cost estimate in this chapter is therefore a lower bound, capturing only the visible fraction of the disease’s economic footprint.
The 2015 Institute of Medicine report established the baseline: 836,000 to 2.5 million Americans affected, with annual economic costs of USD 17–24 billion, including USD 9.1 billion in lost household and labor-force productivity (Committee on the Diagnostic Criteria for Myalgic Encephalomyelitis/Chronic Fatigue Syndrome 2015). This represented USD 18,000–29,000 per patient per year, with indirect costs (lost earnings, reduced work hours, informal care) accounting for approximately two-thirds of the total. The IOM estimate remains the most widely cited single number for U.S. ME/CFS economic burden, though it is a modeled extrapolation from prevalence surveys and cost data, not a direct measurement from claims or administrative databases. (Certainty: 0.77 — government report by National Academies; modeled estimate, not direct cost measurement; extrapolated from survey data. IOM + Clayton 2015 summary share the same evidence base — treated as one source.)
Consequence: The USD 17–24 billion figure — even treated as a lower bound — places ME/CFS in the same economic tier as well-recognized chronic diseases. For a policymaker deciding where to allocate research funding, this number should trigger the same institutional response as the economic burden of multiple sclerosis or rheumatoid arthritis. It has not. Severity applicability: the cost estimate does not stratify by severity; indirect costs are likely underestimated for severe/very severe patients whose caregivers face the highest economic burden.
Australian cost-of-illness studies independently confirm the IOM magnitude and extend it with direct patient-level measurement. Zhao et al. (2023) estimated AU$14.5 billion in total annual costs (AU$28,800 per patient), with indirect costs dominating — lost productivity, informal care, and reduced workforce participation (Zhao et al. 2023). Close et al. (2020), using a separate Australian cohort (n=485), found mean annual costs of AU$14,523 per patient, with 73% attributable to indirect costs and only 16.6% of patients employed (Close et al. 2020). The convergence of two independent Australian studies — different samples, different years, different methodologies — with the IOM U.S. estimate suggests the economic burden is not a methodological artefact of any single study. (Certainty: 0.70 — two independent cost-of-illness studies with convergent estimates; both cross-sectional; n=128 + n=485; Australian data, not U.S. claims-based.)
Consequence: Cross-national replication of the cost magnitude strengthens the case that ME/CFS is inherently expensive — not expensive because of any country’s specific healthcare or disability system. The disease’s cost structure (indirect costs >> direct costs) is consistent across countries, consistent with a disease that removes people from the workforce at high rates and provides them with few effective medical treatments. Severity applicability: neither Australian study stratified costs by severity; the 16.6% employment rate may not generalize to mild patients who continue working part-time.
The KCE Belgian ME/CFS needs assessment (Cornelis et al. 2026, n=749 survey + 19 qualitative interviews) provides the most recent population-level economic data from a government HTA agency. Mean duration of work invalidity: 8.6 years. 73.2% of patients had at least one comorbidity. Diagnostic delay exceeded 2 years for 50% of patients (Cornelis et al. 2026b) (Cornelis et al. 2026a). The KCE report is significant not for any single number but because a government health technology assessment body — whose output typically informs reimbursement and care-pathway decisions — conducted a formal needs assessment for ME/CFS, implicitly recognizing it as a condition warranting systematic health-economic evaluation. (Certainty: 0.68 — government HTA report; survey + qualitative; Belgium-specific; cross-sectional.)
Consequence: When a government HTA agency formally evaluates ME/CFS, it signals institutional recognition that the disease is a legitimate object of health-economic analysis. The 8.6-year mean work invalidity duration translates directly to lost tax revenue, increased disability payments, and reduced household economic security — consequences that accrue to the state, not just the patient. Severity applicability: KCE weighted toward moderate/severe patients (survey methodology); mild patients underrepresented; very severe patients — who cannot complete surveys — entirely absent.
The largest and most recent national estimate comes from Germany, where Risklayer and the ME/CFS Research Foundation modeled the combined societal cost of Long COVID and ME/CFS at EUR 318.8 billion over 2020–2025, including EUR 64.4 billion in 2025 alone — equal to 1.44% of German GDP (Daniell et al. 2026). The model estimated 756,808 people living with Long COVID and 656,951 with ME/CFS at the end of 2025 (over 1.4 million combined), with ME/CFS case numbers still rising while Long COVID declined, and a ME/CFS recovery rate of only ~5% per year — the single factor the authors identify as the largest driver of cost (Daniell et al. 2025). A peer-reviewed German nationwide inpatient analysis confirms the direct-cost component independently: Walter et al. (2022) documented annual direct healthcare costs for Post-COVID Syndrome from claims data (Walter et al. 2022). The German estimate is the first to model both conditions jointly, and its prevalence figures rest on a corrected SARS-CoV-2 infection reconstruction (13–15 million infections in 2025, 80–200 times the official count) that the authors publish as open-source code and data (GmbH 2025). Both the headline cost and the prevalence figures inherit this single correction: it is the load-bearing assumption of the entire estimate, and no independent source — seroprevalence, wastewater surveillance, or claims-based measurement — has yet confirmed an 80–200-fold undercount. The valuation method carries a further systematic bias: human-capital costing prices productivity by wages, which undervalues the unemployed, retired, young, and caregiver populations — the very groups most affected — so the reported figure is likely an underestimate for exactly those people. The resulting prevalence estimates bear on the epidemiology discussion in Epidemiological and Outcomes Research. (Certainty: 0.55 — gray-literature report from a risk-modelling firm and a research foundation with an international scientific advisory board; modeled via Monte Carlo simulation, not directly measured; the authors state the estimate is conservative and likely an underestimate because non-COVID ME/CFS case numbers are held constant. The single-load-bearing-infection-correction risk and the human-capital undervaluation bias are the two principal reasons this figure must be treated as a modeled hypothesis, not a measured fact. Population weight 1.00 — ME/CFS and Long COVID cohort.)
Consequence: A figure of 1.44% of a G7 nation’s GDP reframes ME/CFS and Long COVID from a marginal clinical problem into a measurable share of a national economy — comparable to an entire industrial sector. For a finance minister or research-funding body, that single percentage converts the disease from a health-policy afterthought into an economic-policy variable. Severity applicability: the cost model does not stratify by severity; the authors note indirect costs are likely underestimated for severe and very severe patients, whose caregivers face the highest economic burden (consistent with Family Economic Impact), and the ~5% recovery rate implies the most severely affected fraction persists rather than resolves.
The German model’s 2025 figures are a point on an ascending trajectory, not a peak. ME/CFS case numbers are still rising while Long COVID declines, and ME/CFS recovery is only ~5% per year (Daniell et al. 2026). For any disease whose steady-state prevalence approximates incidence divided by recovery rate, a positive ME/CFS incidence against a ~5% recovery places the eventual steady state far above the current 656,951 cases — so the annual societal cost continues to climb for years even if new SARS-CoV-2-driven conversions stopped today. This is a structural property of the model’s own inputs, not an additional empirical claim. Falsifiable: a subsequent report update showing flat or declining ME/CFS prevalence — with the ~5%-per-year recovery assumption unchanged — would refute the rising-trajectory claim; continued year-over-year prevalence growth would confirm it. (Certainty: 0.40 — arithmetic consequence of the report’s published incidence and ~5%/yr recovery inputs; conditional on the load-bearing infection correction being approximately right; the steady-state framing is a demographic identity, but its magnitude depends entirely on the modeled inputs. Origin: brainstorm.)
Consequence: The economic bill keeps growing for years even if COVID-19 stopped causing new illness today. For a funding body deciding whether to act now, this removes the most convenient argument for delay — that the problem is past its peak and will resolve itself. The rising-prevalence trajectory is precisely why a one-time funding correction (rather than sustained funding) would under-deliver. Severity applicability: the growing steady state is driven by low recovery, which concentrates in the moderate-to-severe fraction that neither fully recovers nor fully participates in the workforce.