A new study published in BMC Health Services Research offers one of the first answers to that question for Taiwan. A team of researchers from the National Defense Medical University and Tri-Service General Hospital, led by corresponding author Ping-Hsuan Hsieh, constructed a formal economic evaluation comparing atogepant with placebo for the preventive treatment of episodic migraine in adults, viewed entirely through the lens of Taiwan’s National Health Insurance system. The work matters because cost-effectiveness is intensely local. A drug that looks like a bargain in one country can look extravagant in another, depending on drug prices, health care utilization patterns, and, crucially, how much a society is willing to pay for an additional year of good health.

The methodological backbone of the study is a Markov model, a standard but powerful tool in health economics. Markov models divide a population into discrete health states and then simulate how individuals move between those states over time, with each transition governed by a probability drawn from clinical trial data. In this case, the researchers defined health states by monthly migraine days, the metric that both clinicians and regulators use to judge whether a preventive therapy is working. Patients cycled through the model in twelve-week intervals over a one-year horizon, a design that mirrors the typical duration of the pivotal trials and keeps the extrapolation burden modest. Treatment discontinuation was built into the model, acknowledging that real-world patients stop preventive drugs far more often than trial participants do.

The clinical engine of the model came from the Phase III ADVANCE trial, the landmark randomized study that established atogepant’s efficacy for episodic migraine. In that trial, patients receiving atogepant experienced meaningful reductions in monthly migraine days compared with placebo, and the transition probabilities derived from those results allowed the Taiwanese team to project how a treated population would distribute across health states over the simulated year. On the cost side, the researchers turned to a source that gives the analysis its local credibility: Taiwan’s National Health Insurance claims data. Rather than importing cost estimates from foreign settings, they extracted real reimbursement figures for outpatient visits, emergency department use, and hospitalization, capturing the actual resource consumption of migraine patients within the system that would ultimately pay for the drug.

The benefits side of the ledger was measured in quality-adjusted life-years, or QALYs, the workhorse metric of cost-utility analysis. A QALY combines survival with health-related quality of life, weighting each year of life by the quality of that year on a scale where one represents perfect health and zero represents death. Migraine rarely kills, which makes the quality-of-life component dominant in this disease area. Patients in the model accrued utility according to their migraine frequency state, with fewer monthly migraine days translating into higher quality of life. Here, however, the researchers confronted a familiar problem in health technology assessment: Taiwan lacks its own utility values for migraine frequency states, so the model had to rely on non-local utility inputs, a limitation that would prove central to interpreting the results.

The base-case findings placed atogepant in an interesting middle zone. Compared with placebo, atogepant generated an incremental cost of 1,766 US dollars and a gain of 0.032 QALYs over the one-year horizon, yielding an incremental cost-effectiveness ratio, or ICER, of 55,749 US dollars per QALY gained. To make sense of that number, health economists conventionally benchmark ICERs against a country’s gross domestic product per capita, a practice popularized by the World Health Organization’s historical guidelines. Taiwan’s GDP per capita stands at roughly 34,040 US dollars, so the atogepant ICER fell between one and two times GDP per capita. Under that framework, the drug is not automatically considered cost-effective at the strictest threshold of one times GDP, but it sits comfortably within the range that many health systems treat as potentially acceptable, particularly for therapies addressing conditions with substantial unmet need.

Because every economic model rests on assumptions that may not hold, the researchers subjected their results to a battery of sensitivity analyses. Deterministic one-way analysis, which varies each parameter individually while holding the rest constant, revealed that the single most influential driver of cost-effectiveness was the utility value assigned to the healthiest state in the model, the state of having zero to three monthly migraine days. This finding is intuitive once unpacked: if avoiding migraines produces only a small quality-of-life improvement, then the QALY gain from an effective preventive drug shrinks, and the cost per QALY balloons. If it produces a large improvement, the opposite occurs. Since that utility value was imported rather than measured locally, the entire cost-effectiveness conclusion hinges on a number that may or may not reflect how Taiwanese patients actually experience migraine freedom.

The probabilistic sensitivity analysis added a second layer of nuance by running the model thousands of times while simultaneously sampling every parameter from its uncertainty distribution. The result was a cost-effectiveness acceptability curve showing how the probability that atogepant is cost-effective changes as society’s willingness to pay rises. At a threshold of one times Taiwan’s GDP per capita, atogepant had only a 2.9 percent chance of being cost-effective. At two times GDP per capita, that probability climbed to 74.4 percent, and at three times GDP per capita it reached 96.0 percent. The tipping point, where the probability crossed fifty percent, came at a willingness to pay of approximately 56,300 US dollars per QALY. In plain terms, whether atogepant is a good buy depends almost entirely on how much value Taiwan’s health system assigns to relieving the burden of episodic migraine, and the drug becomes the more likely-than-not choice just above the two-times-GDP threshold.

The authors are appropriately candid about the limits of their work, describing the study as a baseline, proof-of-concept assessment rather than a definitive verdict. Beyond the reliance on non-local utility values, the comparison against placebo, while methodologically clean, sidesteps the question that payers ultimately care about most: how atogepant performs against the preventive therapies already in use, from older oral agents to the injectable CGRP monoclonal antibodies. The one-year horizon, chosen to align with trial evidence, also cannot capture longer-term effects, and the model did not incorporate indirect costs such as lost workplace productivity, which for migraine are often larger than direct medical costs. The researchers call for future evaluations to integrate Taiwan-specific quality-of-life data and to employ network meta-analyses positioning atogepant against active comparators.

Even with those caveats, the study marks a meaningful step for evidence-based drug policy in Asia. Taiwan’s National Health Insurance is renowned for delivering universal coverage at remarkably low administrative cost, but that efficiency depends on disciplined decisions about which new technologies to adopt. By grounding an atogepant evaluation in local claims data and transparent GDP-based thresholds, the research team has created a template that other middle- and high-income health systems can adapt, while simultaneously flagging exactly where local data collection, particularly patient-reported quality of life across migraine frequency states, would deliver the greatest payoff. For the millions of people whose migraines could be prevented by CGRP-targeted therapy, the message is one of cautious optimism: the science works, the economics are within reach, and the remaining uncertainty is now precisely mapped for the next generation of studies to resolve.

Subject of Research: Cost-effectiveness evaluation of atogepant versus placebo for preventive treatment of episodic migraine in Taiwan

Article Title: Economic evaluation of atogepant compared to placebo for the preventive treatment of episodic migraine in Taiwan

Article References: Fang, C.-Y., Lin, T.-K., Yang, F.-C., Liu, Y., Ruan, Y.-Z., & Hsieh, P.-H. (2026). Economic evaluation of atogepant compared to placebo for the preventive treatment of episodic migraine in Taiwan. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15776-5

Image Credits: AI Generated

DOI: 10.1186/s12913-026-15776-5

Keywords: atogepant, CGRP antagonist, episodic migraine, cost-effectiveness, health economics, QALY, ICER, Markov model, Taiwan National Health Insurance, migraine prevention, pharmacoeconomics, health technology assessment

Acquired hemolytic anemia, or AHA, is not a single disease but a heterogeneous collection of conditions in which red blood cells are destroyed prematurely by mechanisms outside the cells themselves. Because the body cannot replace the lost cells quickly enough, patients are left with too few oxygen-carrying red cells, producing dizziness, weakness, irregular heartbeat, and shortness of breath. In severe cases, free hemoglobin spilled into the bloodstream can trigger kidney injury, stroke, or heart attack, and the disorder’s causes range from autoimmune attack and transfusion reactions to medications, infections, and metal poisoning. The most common subtype worldwide, autoimmune hemolytic anemia, strikes roughly one to three people per 100,000 each year, while paroxysmal nocturnal hemoglobinuria affects only about 15.9 individuals per million.

Standard treatment depends on the underlying cause and includes corticosteroids, B-cell-depleting monoclonal antibodies, blood transfusion, splenectomy, or hematopoietic stem cell transplantation. Yet even with these options, patients remain vulnerable to serious complications, including cardiovascular disease, immune dysfunction, infection, thrombotic events, and pulmonary hypertension. That persistent burden of illness is precisely why the Taiwanese research team, led by investigators affiliated with China Medical University, set out to ask whether Chinese herbal medicine, which is reimbursed and systematically recorded under Taiwan’s National Health Insurance program, might be associated with better long-term outcomes in this population.

The team mined claims data from 2008 to 2019, identifying patients newly diagnosed with acquired hemolytic anemia between 2010 and 2017 using standard diagnostic codes. To ensure diagnoses were reliable, they required at least three outpatient claims or one inpatient claim for the condition. After excluding patients missing demographic information, those who underwent splenectomy, and those whose herbal prescriptions fell below a prespecified cumulative threshold of 14 days in the year following diagnosis, the cohort comprised 6,807 eligible patients, of whom 909 were classified as herbal medicine users and 2,229 as non-users.

Because patients who seek out traditional medicine often differ systematically from those who do not, the researchers used propensity score matching to build comparable groups. Matching on age, sex, Charlson comorbidity index, blood transfusion history, and iron chelation therapy yielded 790 well-balanced pairs. Before matching, herbal users were younger, more likely to be female, and carried a lower comorbidity burden than non-users, differences that the matching procedure largely erased and that the subsequent statistical models further adjusted for.

The survival signal was striking. In a fully adjusted Cox proportional hazards model, herbal medicine users had a 37 percent lower risk of death from any cause than matched non-users, with an adjusted hazard ratio of 0.63 and a 95 percent confidence interval of 0.51 to 0.77. Kaplan-Meier curves confirmed better survival among users, with a log-rank p-value of 0.0006. The benefit extended to the cardiovascular realm: after multivariable adjustment, herbal treatment was associated with a 62 percent reduction in cardiovascular disease-related mortality, corresponding to an adjusted hazard ratio of 0.38. Older age, heavier comorbidity burden, and blood transfusion each independently raised mortality risk, while female sex was linked to lower all-cause mortality. Sensitivity analyses using different cumulative exposure cutoffs produced consistent results.

Beyond the survival analysis, the study broke new ground by mapping how physicians actually combine herbal products in this patient group. Using association rule mining and network visualization, the team analyzed more than 27,000 prescriptions and identified two distinct clusters. The first, a hematologic-supportive cluster, centered on Dang-Gui, Huang-Qi, Gui-Pi-Tang, E-Jiao, and Dan-Shen, with the Dang-Gui to Huang-Qi pair showing the strongest co-prescription relationship. The second cluster paired Jie-Geng with Bei-Mu. Gui-Pi-Tang was the most frequently prescribed formula overall, while Huang-Qi, Dan-Shen, Bei-Mu, Jie-Geng, Dang-Gui, and E-Jiao led among single herbs.

Experimental literature offers plausible biological explanations for why these particular herbs cluster together. Polysaccharides from Angelica sinensis, the botanical source of Dang-Gui, appear to protect hematopoietic cells from oxidative injury and reverse anemia of chronic disease in animal models. Huang-Qi, derived from Astragalus species, has been shown to raise red cell, white cell, hemoglobin, and hematocrit levels in anemic animals, and its principal constituent astragaloside IV has demonstrated anti-anemic effects in chemotherapy-induced models as well as potential benefits in heart failure. Dan-Shen, the root of Salvia miltiorrhiza, reduces erythrocyte membrane damage and improves blood viscosity, while collagen-derived peptides from E-Jiao stimulate hematopoiesis. Jie-Geng, meanwhile, has shown cardioprotective effects by blocking angiotensin II-induced cardiomyocyte apoptosis. These mechanisms align neatly with the observed reduction in both overall and cardiovascular mortality.

The findings echo earlier Taiwanese studies in related blood disorders. Prior work found that herbal medicine use was associated with improved survival in aplastic anemia and with reduced all-cause, diabetes-related, and cardiovascular mortality in hereditary hemolytic anemias, with overlapping prescription patterns featuring Dang-Gui, Huang-Qi, and Dan-Shen. The consistency across these independent cohorts strengthens the case that the current results are not a statistical fluke, though it also means the new study inherits some of the same methodological constraints.

Those constraints matter. Claims data contain no laboratory values or physical assessments, leaving room for residual confounding and possible misclassification of exposure or outcome. The excluded patients differed from the included ones in sex, age, comorbidity, and transfusion use, raising the possibility of selection bias. Individualized dosing information was unavailable, and the matching design precluded dose-response analysis. The authors also note that multi-herb combinations common in traditional practice may interact with pharmaceuticals in ways the study could not capture, and that healthcare systems elsewhere may differ from Taiwan’s. Still, the researchers argue that the identified herbal clusters provide a concrete, testable roadmap for prospective cohort studies and mechanistic experiments aimed at determining whether these centuries-old formulations can genuinely extend survival for patients with acquired hemolytic anemia.

Subject of Research: Association between Chinese herbal medicine use and mortality in acquired hemolytic anemia

Article Title: Chinese herbal medicine therapy reduces all-cause and cardiovascular disease-related mortalities in acquired hemolytic anemia: A Taiwanese nationwide retrospective cohort study

Article References: Chou, C.-H., Chiou, J.-S., Hsu, Y.-L., Chiu, M.-L., Chen, C.-J., Liang, W.-M., Tsai, F.-J., Wu, Y.-C., Chou, I.-C., Lin, T.-H., Liao, C.-C., Huang, S.-M., Lin, C.-W., Li, T.-M., Li, J.-P., & Lin, Y.-J. (2026). Chinese herbal medicine therapy reduces all-cause and cardiovascular disease-related mortalities in acquired hemolytic anemia: A Taiwanese nationwide retrospective cohort study. Heliyon, 12(15), Article e45157. https://doi.org/10.1016/j.heliyon.2026.e45157

Image Credits: AI Generated

DOI: 10.1016/j.heliyon.2026.e45157

Keywords: acquired hemolytic anemia, Chinese herbal medicine, cardiovascular mortality, all-cause mortality, Taiwan National Health Insurance, propensity score matching, Cox proportional hazards, Dang-Gui, Huang-Qi, autoimmune hemolytic anemia, prescription pattern analysis, traditional Chinese medicine