Dengue Deaths Hide in Unconfirmed Cases, 13-Year Philippine Surveillance Study Reveals

A sweeping 13-year analysis of dengue surveillance data from Davao City, Philippines, has revealed a striking paradox at the heart of routine disease monitoring: while laboratory-confirmed dengue cases rose significantly over time, the vast majority of recorded dengue deaths occurred among patients who were never laboratory-confirmed at all. The study, published in Frontiers in Tropical Diseases, examined 81,075 dengue notifications reported through the Philippine Integrated Disease Surveillance and Response (PIDSR) system between January 2012 and December 2024, and found that the three surveillance case classifications—suspected, probable, and confirmed—each told a distinctly different story about the epidemic unfolding in this major southern Philippine city.

Dengue remains one of the most formidable public health challenges in tropical and subtropical regions. The mosquito-borne viral disease ranges in clinical presentation from a self-limiting febrile illness to severe disease marked by plasma leakage, hemorrhage, organ dysfunction, and death. In 2024 alone, more than 13 million dengue cases and over 8,700 dengue-related deaths were reported globally. The Philippines ranks among the most affected countries in Southeast Asia, where cyclical outbreaks continue to strain healthcare and public health systems. Yet interpreting long-term dengue trends has always been complicated by changes in case definitions, reporting practices, and the availability of laboratory confirmation—factors that can reshape the apparent epidemiological landscape without any true change in transmission.

The research team, led by Alfredo A. Hinay Jr. of the University of the Immaculate Conception in Davao, conducted a retrospective observational study using the official citywide surveillance database coordinated by the City Epidemiology and Surveillance Unit. In the Philippine system, dengue is a nationally notifiable disease, and healthcare providers must report cases meeting national surveillance definitions. The three-tier classification system distinguishes suspected cases—previously healthy individuals with acute febrile illness of two to seven days duration and clinical signs consistent with dengue—from probable cases, which additionally show a positive dengue IgM antibody test, and confirmed cases, which require positive results from viral culture isolation, polymerase chain reaction (PCR), or dengue NS1 antigen testing. Each tier represents a different level of diagnostic certainty, and each captured a different slice of the epidemic.

Of the 81,075 notifications analyzed after data cleaning, 59,021 were classified as suspected cases, 20,559 as probable cases, and only 1,495 as confirmed cases. Suspected cases therefore dominated the surveillance record throughout the entire 13-year period, accounting for nearly three-quarters of all notifications. Demographically, the burden fell hardest on children: individuals younger than 15 years represented 59.4 percent of suspected cases, 61.4 percent of probable cases, and 62.7 percent of confirmed cases. Males comprised a slightly higher proportion of notifications than females across all classifications, ranging from 51.2 to 52.2 percent. Hospital admission was reported in more than 90 percent of notifications across every classification, indicating that patients classified at all levels of diagnostic certainty received comparable clinical attention within the healthcare system.

The temporal analysis, performed using the Mann–Kendall trend test—a non-parametric method for detecting monotonic trends in time-series data that does not require assumptions of normality—produced the study’s most consequential finding. Confirmed dengue incidence increased significantly over the study period (p = 0.003), while probable dengue incidence remained stable with no significant trend (p = 0.855), and suspected dengue incidence showed a decreasing, though not statistically significant, tendency (p = 0.059). In other words, the same city, the same virus, and the same surveillance system produced three divergent trend lines depending solely on how cases were classified. Had the analysis been restricted to laboratory-confirmed cases alone, observers would have concluded that dengue was steadily worsening; had it relied on suspected cases, the conclusion might have been one of gradual improvement.

The increasing trend in confirmed dengue incidence proved remarkably consistent across every stratum examined. All three congressional districts of Davao City showed significant upward trends—Congressional District I (p = 0.005), Congressional District II (p = 0.016), and Congressional District III (p = 0.002)—and the increase was significant among both children under 15 (p = 0.005) and adults aged 15 or older (p = 0.014), as well as among both females (p = 0.005) and males (p = 0.003). This uniformity suggests that the rise in confirmed cases was not confined to particular neighborhoods or demographic groups but reflected a broad, citywide phenomenon. The authors attribute such increases in other dengue-endemic settings to a combination of changing transmission dynamics, expanded availability of diagnostic testing, improvements in laboratory capacity, and enhanced surveillance practices. Notably, the Department of Health only formally recognized positive NS1 antigen test results as a criterion for laboratory-confirmed dengue in 2024, so its influence on the long-term trend is likely limited, though it may have shaped the distribution of classifications in the final year.

The COVID-19 pandemic casts a long shadow over any interpretation of these trends. Consistent with reports from other dengue-endemic countries, dengue notifications in Davao City declined during the pandemic, likely reflecting movement restrictions, school closures, altered healthcare-seeking behavior, and disruptions to routine surveillance and healthcare services. The subsequent increase following the relaxation of public health measures may represent a post-pandemic rebound driven by resumed population mobility, renewed opportunities for virus transmission, and improved case detection. However, the authors emphasize that the increasing trend in laboratory-confirmed dengue remained evident beyond the immediate post-pandemic period, suggesting the pattern cannot be explained solely by pandemic effects.

Perhaps the most sobering finding concerns mortality. Dengue mortality incidence showed no significant temporal trends in any classification—suspected (p = 0.089), probable (p = 0.242), or confirmed (p = 0.282)—yet the distribution of deaths across classifications was deeply asymmetric. Probable cases exhibited the highest cumulative case fatality rate at 0.85 percent (rising to 0.89 percent among admitted patients), followed by confirmed cases at 0.67 percent, and suspected cases at just 0.26 percent. Because suspected and probable cases vastly outnumbered confirmed ones, they accounted for the overwhelming majority of deaths recorded among reported dengue notifications. Among non-admitted patients, case fatality rates were low across all classifications, and no deaths were recorded among non-admitted confirmed notifications. The pattern suggests that diagnostic certainty and clinical severity do not necessarily align: patients may die of clinically compatible dengue without ever receiving the laboratory test that would place them in the confirmed category.

The authors caution that routine passive surveillance systems carry inherent biases, including underreporting, delayed reporting, and misclassification. Patients with more severe presentations may have been more likely to undergo confirmatory testing and thus be classified as probable or confirmed, while those with milder illness or limited access to laboratory services may have remained classified as suspected. The study also could not clinically validate deaths recorded within the surveillance system, evaluate alternative diagnoses, or determine the extent to which individual deaths were directly attributable to dengue infection. Without an external reference standard, surveillance performance measures such as sensitivity and positive predictive value could not be assessed. These limitations temper but do not erase the central conclusion.

For low- and middle-income countries, where dengue surveillance systems frequently operate under constraints related to diagnostic availability and laboratory capacity, the implications are substantial. Analyses restricted exclusively to confirmed cases may underestimate severe outcomes recorded within routine surveillance systems, since probable and suspected cases accounted for most deaths despite stable incidence trends over time. The authors argue that surveillance classifications beyond laboratory-confirmed dengue continue to provide meaningful epidemiological information about disease burden and severe outcomes in resource-constrained settings. As dengue continues its global expansion driven by urbanization, population growth, climate variability, and expanding mosquito habitats, the Davao City experience offers a cautionary lesson: the picture of an epidemic depends profoundly on where in the diagnostic pipeline one chooses to look, and public health decisions built on a single classification may miss the deaths that matter most.

Subject of Research: Temporal trends in dengue incidence and mortality across suspected, probable, and confirmed surveillance case classifications in Davao City, Philippines

Article Title: Temporal trends in dengue incidence and mortality across surveillance case classifications: a 13-year analysis from Davao City, Philippines

Article References: Hinay, A. A., Jr, Reyes, J. A. H., & Plaza, G. C. G. (2026). Temporal trends in dengue incidence and mortality across surveillance case classifications: a 13-year analysis from Davao City, Philippines. Frontiers in Tropical Diseases, 7, Article 1899788. https://doi.org/10.3389/fitd.2026.1899788

Image Credits: AI Generated

DOI: 10.3389/fitd.2026.1899788

Keywords: dengue, surveillance, Philippines, epidemiology, mortality, case classification, public health, Mann-Kendall trend test, Davao City, PIDSR, low- and middle-income countries, vector-borne disease