The research team, led by Namita Kaushik and Andreas Tietz, screened 6,983 records and identified 341 studies reporting healthcare resource utilisation or economic outcomes in dengue. Of these, 44 studies reporting hospitalisation rates and 121 studies reporting length of stay met the rigorous feasibility criteria for quantitative synthesis. The work followed PRISMA guidelines, was registered prospectively in the PROSPERO database, and applied the Downs and Black checklist to appraise study quality. Data were drawn from five biomedical databases searched via the Ovid platform, supplemented by citation searches and conference abstracts, with no restriction on country of origin.

The headline result is a pooled global hospitalisation rate of 36.7 percent (95 percent confidence interval 29.0 to 44.4) among patients with dengue irrespective of severity, derived under a random-effects model. The mean length of stay across hospitalised patients was 4.9 days (95 percent CI 4.6 to 5.2). Both outcomes showed very high statistical heterogeneity, with I-squared values approaching 100 percent, reflecting the enormous diversity of settings, designs, and populations represented in the underlying literature. To understand that variation, the authors performed meta-regression across a range of study-level covariates, including geography, outbreak status, data source, and study design.

One variable dominated all others: disease severity. Severity was significantly associated with both higher hospitalisation rates and longer stays, explaining 31.4 percent of between-study variance in hospitalisation and 11.4 percent of the variance in length of stay, with p-values below 0.0001 in both models. Subgroup analyses quantified the gradient precisely. Patients with non-severe dengue without warning signs were hospitalised at a pooled rate of 17.7 percent, rising to 36.9 percent once WHO warning signs such as persistent vomiting, abdominal pain, mucosal bleeding, raised haematocrit, or thrombocytopaenia appeared. Among patients classified as having severe dengue, the pooled hospitalisation rate reached 79.4 percent.

Length of stay followed the same severity gradient. Patients with non-severe disease averaged 4.6 days in hospital, compared with 6.5 days for severe dengue. The most dramatic figures emerged when complications were considered: patients with severe dengue or dengue haemorrhagic fever who developed complications such as shock, renal impairment, platelet-transfusion-requiring thrombocytopaenia, or sepsis stayed a mean of 15.5 days, roughly double the 7.5 days observed in severe cases without complications. Co-infections with pathogens such as SARS-CoV-2 or chikungunya virus, and comorbidities including haematologic malignancies, autoimmune disease, and obesity, each added roughly a day or more to the average stay.

Regional patterns add a further layer of complexity. Hospitalisation rates were numerically higher in Asia, at 35.1 percent, than in Latin America, at 29.3 percent, and length of stay was correspondingly longer in Asian settings, at 5.1 days versus 4.1 days in Latin America. The authors note that this Asian excess echoes the 2016 analysis by Shepard and colleagues, which reported rates of 19.3 to 24.8 percent in Asian subregions against just 3.8 percent in Latin America and the Caribbean. Possible explanations include differing admission thresholds, a higher force of infection, multiple co-circulating serotypes, and demographic differences, but the authors emphasise that the drivers remain unclear and warrant targeted investigation.

The timing of the analysis is significant. Since the last global estimates were published, the world has witnessed record-breaking dengue waves, shifts in circulating serotypes, changing patient demographics, the disruption of the COVID-19 pandemic, and improvements in surveillance and clinical preparedness. The World Health Organization reported a doubling of reported cases from around 7 million in 2023 to approximately 14 million in 2024, while modelling work published in 2025 estimated that annual apparent cases rose from 26.45 million in 1990 to 58.96 million in 2021. Against that epidemiological backdrop, a hospitalisation rate approaching 37 percent implies a demand for inpatient beds, nursing capacity, and supportive care that many endemic health systems are ill-equipped to absorb during peak transmission seasons.

Methodological sensitivity analyses strengthen the central conclusions while exposing their limits. Excluding studies with fewer than 100 participants and those with more than 50,000 yielded a pooled hospitalisation rate of 29.4 percent with narrower confidence intervals, and excluding studies whose data were collected during the pandemic produced an estimate of 25.3 percent, hinting that overwhelmed COVID-era hospitals may have temporarily suppressed dengue admissions. A single exceptionally large nationwide Brazilian cohort dominated the fixed-effects estimate, but random-effects modelling, which the authors favoured given the heterogeneity, attenuated its influence. For length of stay, no evidence of publication bias was detected by Egger’s regression, and the GRADE certainty of evidence was rated low, primarily because of the observational design of the included studies rather than inconsistency or imprecision.

The findings carry immediate practical implications. Hospitalisation is recognised as a key outcome in dengue research: an international Delphi consensus involving experts from more than 30 countries identified the need for hospitalisation as a core endpoint for patients and clinicians. The new benchmarks should therefore help clinical trialists select endpoints and calculate sample sizes, and they align with recently published WHO target product profiles that recommend dengue-related hospitalisation as a secondary efficacy outcome in non-severe disease. For policymakers, the estimates provide a foundation for cost-effectiveness analyses of vaccines, antivirals, and vector-control strategies, and for identifying the regions where preventive investment would relieve the greatest pressure on hospitals.

The authors are careful to frame the pooled figures as average trends across highly diverse settings rather than universal constants. Admission thresholds, clinical practice, healthcare-seeking behaviour, and reporting conventions differ substantially between countries, and the denominator in most studies was skewed towards patients already seeking care, which may inflate the apparent hospitalisation rate relative to all symptomatic infections. The observation that some patients with warning signs or severe disease were nonetheless not hospitalised may reflect symptom misclassification, resource constraints, or emergency department stays that were never recorded as admissions. Even so, the overall message is difficult to escape: dengue’s demand on hospital infrastructure has likely risen considerably over the past decade, and quantifying the associated direct and indirect costs, productivity losses, and caregiver burden is now an urgent priority for global health planning.

Subject of Research: Global hospitalisation rates and length of stay for dengue patients

Article Title: Global burden of dengue on healthcare systems: a meta-analysis of healthcare resource utilisation

Article References: Kaushik, N., Subramanian, S., Gedif, K., Nallagangula, T. K., Singuru, S., Panchal, S., Meyer-Andrieux, I., Gruenberger, J.-B., McBride, A., Yacoub, S., & Tietz, A. (2026). Global burden of dengue on healthcare systems: a meta-analysis of healthcare resource utilisation. eClinicalMedicine, 100, Article 104217. https://doi.org/10.1016/j.eclinm.2026.104217

Image Credits: AI Generated

DOI: 10.1016/j.eclinm.2026.104217

Keywords: dengue, hospitalisation rate, length of stay, meta-analysis, healthcare resource utilisation, systematic review, global health burden, WHO warning signs, severe dengue, Aedes mosquitoes, health systems, eClinicalMedicine