Abstract
Objectives:
Puerto Rico’s 2024-2025 dengue epidemic highlighted the need to understand how complementary surveillance systems capture cases and severity. We compared the Sentinel Enhanced Dengue Surveillance System (SEDSS) and the Passive Arboviral Disease Surveillance System (PADSS) in capturing characteristics of dengue cases during this epidemic and assessed their complementary roles in epidemic monitoring and public health preparedness.
Methods:
We analyzed laboratory-confirmed dengue cases reported in SEDSS and PADSS from January 1, 2024, through January 31, 2025. SEDSS recruits at sentinel sites, collecting clinical, epidemiological, and laboratory data, while PADSS relies on clinician-initiated reporting across the island. We used descriptive statistics, cross-correlation analyses, and generalized additive models to compare temporal trends, demographic characteristics, clinical features, and severe dengue outcomes.
Results:
SEDSS enrolled 373 dengue patients (7.0% of tested patients), while PADSS reported 6488 dengue patients (60.1% of tested patients). Both systems showed aligned epidemic peaks, although PADSS detected more cases overall. Compared with PADSS patients, SEDSS patients were younger (median age = 22 vs 27 y) and had higher proportions of warning signs, including mucosal bleeding (21.7% vs 6.9%), hemoconcentration (4.4% vs 0.1%), and restlessness (31.1% vs 7.5%) (P < .001 for all). Severe dengue was more common in SEDSS patients (9.1% vs 5.6%; P = .02), likely due to more detailed clinical data, with the highest rates among patients aged 10 to 19 years (16.3%) and <10 years (10.5%). SEDSS captured severe plasma leakage (6.2%), which was not recorded in PADSS. PADSS provided broader geographic coverage.
Conclusions:
SEDSS captures detailed clinical data, whereas PADSS provides broader coverage and higher case counts. Integrating both systems strengthens epidemic response, resource allocation, and public health decision-making.
Dengue, caused by 1 of 4 dengue viruses (DENV-1 through DENV-4), is transmitted primarily by Aedes species mosquitoes.1,2 Dengue remains a global health challenge, causing approximately 105 million infections each year. 3 The disease ranges from mild febrile illness to severe dengue, characterized by severe plasma leakage, severe bleeding, or organ impairment. DENV infections lead to substantial morbidity and mortality, particularly among children and adolescents aged <15 years and people with certain underlying conditions.4-6 Control efforts—vector management, public health interventions, and vaccination—face challenges due to vector adaptability, limited prevention tools, fluctuating serotype circulation, and climate variability. 7
Dengue is endemic in Puerto Rico, with the last outbreak reported in 2012-2013. In 2024, an outbreak—driven by a shift from DENV-1 to DENV-2 and DENV-3—prompted the Puerto Rico Department of Health (PRDH) to declare an emergency.8-10 This resurgence reflects the dynamic nature of dengue serotype circulation and the need for surveillance systems to detect epidemics, monitor transmission, and support timely public health action.
Puerto Rico uses 2 complementary systems to monitor arboviral diseases, including dengue: the Passive Arboviral Disease Surveillance System (PADSS) and the Sentinel Enhanced Dengue Surveillance System (SEDSS). PADSS relies on clinicians to report suspected cases and submit specimens for laboratory testing. The system captures a broad geographic range but potentially underreports cases when testing is not performed or when reports lack clinical detail, which is not always required. In contrast, SEDSS actively recruits febrile patients at selected health care facilities and collects standardized clinical, epidemiological, and laboratory data. 11
PADSS and SEDSS differ in case capture methods, data completeness, and geographic coverage, but their performance has not been directly compared during a large-scale dengue epidemic. This analysis evaluated how PADSS and SEDSS differ in the demographic and clinical characteristics of dengue case detection, identification of severe disease, and reflection of epidemic trends. Understanding these differences can inform surveillance strategies, improve epidemic detection, and guide resource allocation. Beyond comparing 2 Puerto Rico–specific systems, our analysis demonstrates how pairing a geographically broad passive system with a clinically rich sentinel system can identify serotype shifts weeks earlier than PADSS alone, quantify severity patterns with higher precision than either system alone, and interpret prolonged, atypical epidemic profiles in real time. Although both systems are conducted in a tropical island setting with year-round dengue transmission, their core principles—rapid laboratory turnaround, routine serotype monitoring, and integrated interpretation across complementary systems—are transferable to surveillance for other arboviruses and emerging infections in both island and continental settings.
Methods
SEDSS
Established in 2012 by the Centers for Disease Control and Prevention (CDC) and Ponce Health Sciences University,11-13 SEDSS was designed to enhance the detection and characterization of arboviral and respiratory infections among patients with acute febrile illnesses. It provides comprehensive demographic, clinical, and laboratory data, enabling more complete disease characterization than passive surveillance.
SEDSS currently operates at 3 sites: Centro Medico Episcopal San Lucas and Centro de Emergencia y Medicina Integrada in Ponce and Auxilio Mutuo Hospital in San Juan. Recruitment is conducted in emergency departments via convenience sampling. Patients of any age reporting fever within the previous 7 days are eligible for enrollment. In April 2020, eligibility expanded to include patients with cough or dyspnea within the previous 14 days, regardless of fever, to improve surveillance of respiratory viruses after the COVID-19 pandemic. Enrollment requires collection of a blood sample and a nasopharyngeal or oropharyngeal swab. 12 Convalescent serum samples are collected for a subset of participants.
SEDSS collects data through patient interviews and medical record reviews. Case investigation forms are used to record demographic characteristics, comorbidities, and clinical features at enrollment. The convalescent sample-processing form documents additional clinical information 7 to 14 days after enrollment. For hospitalized patients, the hospital admitted abstraction form captures detailed clinical indicators of disease severity and progression.
SEDSS cases are reported to PRDH as part of routine surveillance. Records are merged in the arboviral surveillance database, with automated and manual checks for duplicate entries based on patient identifiers (eg, name, date of birth, specimen collection date). When a case appeared in both systems, we retained the SEDSS record in the SEDSS dataset and excluded it in PADSS for this analysis to maintain exclusivity between the 2 systems.
Institutional review boards at CDC, Auxilio Mutuo Hospital, and Ponce Medical School Foundation approved the study. CDC approval was under protocol 6214; local institutional approvals were under protocol number 120308-VR and, subsequently, protocol number 2311173707. All adult participants (aged ≥21 y) and emancipated minors provided written consent. For nonemancipated minors aged 14 to 20 years, a parent/guardian provided written consent and the participant provided written assent. For children and adolescents aged 7 to 13 years, a parent/guardian provided written consent and the participant provided assent.
PADSS
PADSS relies on (1) mandated reporting of suspected arboviral infections from clinicians who send specimens to PRDH for diagnostic testing and (2) mandated reporting of positive test results from private laboratories across Puerto Rico. 14 It captures data on symptom onset, case location, and outcomes such as hospitalization and death. For detailed clinical information, PADSS depends on patient interview (when available) or optional entry by health care providers; the system does not provide comprehensive data for all cases. Although PADSS also surveys other reportable arboviruses such as Zika and chikungunya viruses, no confirmed cases of these infections were reported during the study period; we included only laboratory-confirmed dengue cases in this analysis.
Study Period
For brevity, we refer to this as the 2024-2025 epidemic period, although transmission remained persistently above the historical epidemic threshold from January 2024 through the end of the study period (January 2025), with cases peaking in October 2024 and continuing above the outbreak threshold through mid-2025.8,10,15
Definitions
We defined confirmed dengue as detection of DENV by positive reverse transcription polymerase chain reaction (RT-PCR) or detection of nonstructural protein 1 (NS1) antigen, and we defined probable dengue as detection of anti-DENV immunoglobulin M (IgM) antibodies using enzyme-linked immunosorbent assay (IgM ELISA) in a patient’s serum specimen. For SEDSS, all specimens collected ≤7 days after symptom onset underwent RT-PCR testing, and specimens collected >3 days after symptom onset were also tested by IgM ELISA. For PADSS, specimens collected at any point after symptom onset were initially tested by RT-PCR. Specimens with negative RT-PCR results were then tested by IgM ELISA. Testing at commercial laboratories included IgM ELISA and NS1 antigen assays. We included both confirmed and probable cases as laboratory-confirmed dengue in the analysis.
We classified dengue warning signs and severe dengue per World Health Organization guidelines, 16 incorporating clinical data from interviews and medical records for SEDSS participants, which were captured in the case investigation form, convalescent sample-processing form, and hospital admitted abstraction form. Warning signs included abdominal pain, persistent vomiting, clinical fluid accumulation, mucosal bleeding, restlessness, hemoconcentration, and hepatomegaly (eTable 1 in the Supplement). We categorized severe dengue as severe plasma leakage, severe bleeding, or severe organ impairment.16,17 In PADSS, health care providers could also indicate severe dengue based on clinical judgment reflecting these criteria.
Statistical Analysis
We compared case detection and temporal trends between the 2 systems by aggregating weekly confirmed and probable cases. We calculated relative differences in cases detected and assessed temporal lags by using cross-correlation analyses. A lag of 0 indicates same-week correlation; however, with a maximum possible value of 1.0, an r of 0.50 reflects only moderate agreement. To complement the cross-correlation analyses, we used cumulative distribution function analysis, which assessed whether the overall timing of case accumulation was similar between systems. A generalized additive model (GAM) (mgcv package in R, 18 quasi-Poisson with log link) modeled weekly SEDSS case counts as a smooth function of weekly PADSS case counts to allow nonlinearity.
We then compared dengue case numbers, demographic and clinical characteristics, and severe dengue indicators between SEDSS and PADSS. We used descriptive statistics to summarize patient characteristics, clinical signs, and disease severity. To determine statistically significant (hereinafter, significant) differences, we conducted Pearson χ2 tests or Fisher exact tests for categorical variables and the Mann–Whitney U test for continuous variables. We used the Benjamini–Hochberg procedure 19 to adjust P values; we set the significance at P < .05. We conducted subanalyses among hospitalized patients. We performed statistical analyses using R software version 4.4.0 (R Foundation for Statistical Computing).
Results
Comparison of Case Detection and Temporal Trends
From January 1, 2024, through January 31, 2025, SEDSS recorded 5642 visits due to acute febrile illness among 5363 participants. Dengue testing was performed in samples from 5321 visits (5074 participants), identifying 373 dengue cases (7.0% positivity rate): 355 (95.2%) confirmed by RT-PCR or NS1 and 18 (4.8%) probable by IgM. Dengue testing results were not available for 321 visits due to inadequate sample quality (38.0%; n = 122) or pending results (62.0%; n = 199) at the time of analysis. During the same period, PADSS recorded 10 801 tested patients, identifying 6488 dengue cases (60.1%): 5020 (77.4%) confirmed by RT-PCR, 334 (5.1%) confirmed by NS1, and 1134 (17.5%) probable by IgM. PADSS included both PRDH testing (9562 patients; 5249 positive [54.9%]) and private laboratories (1239 patients; 1239 positive [100%, only positives reported]). Time-series plots (Figure 1A) demonstrated aligned peaks in dengue activity across both systems (confirmed and probable cases) until December 2024, when SEDSS case detection declined while PADSS remained stable. The PADSS-to-SEDSS case ratio varied widely during the study period (median [IQR]: 16.2 [9.9-24.2]) but narrowed during peak transmission, indicating closer alignment when dengue activity was highest (Figure 1B).
(A) Time-series plot of Passive Arboviral Disease Surveillance System (PADSS) and Sentinel Enhanced Dengue Surveillance System (SEDSS) confirmed and probable cases of dengue by week. (B) Winsorized relative difference in cases detected (PADSS vs SEDSS), Puerto Rico, January 1, 2024–January 31, 2025. In B, bars indicate the weekly ratio of PADSS to SEDSS dengue cases, Winsorized at the 5th and 95th percentiles to limit the influence of extreme values. The dashed line indicates a locally estimated scatterplot smoothing trend line, showing smoothed temporal variation in weekly relative differences between PADSS and SEDSS cases. Shading indicates 95% CIs. The median (IQR) weekly PADSS-to-SEDSS case ratio was 16.2 (9.9-24.2), and the mean was 23.8, reflecting higher overall case detection in PADSS than in SEDSS. Weekly ratios varied with epidemic intensity: systems were more closely aligned during high-transmission weeks, as further explored (Figure 2). Data sources: PADSS 14 ; SEDSS.11-13
GAM analysis (Figure 2) revealed a nonlinear but generally increasing relationship between weekly PADSS and SEDSS case counts. When PADSS reported low weekly case totals (eg, <100 cases), the GAM predicted approximately 5 SEDSS cases per week, whereas SEDSS detections rose steadily with increasing PADSS counts. The smooth term was significant (P = .007), and the model explained 31.4% of deviance.
Generalized additive model (GAM) of weekly cases of dengue identified by the Passive Arboviral Disease Surveillance System (PADSS) and Sentinel Enhanced Dengue Surveillance System (SEDSS), Puerto Rico, January 2024–January 2025. The solid line depicts the GAM-predicted number of SEDSS cases based on PADSS case counts, using a smooth function to capture nonlinear trends (quasi-Poisson with log link). Shading indicates 95% CIs; points show observed weekly values. When PADSS reported <100 cases per week, predicted SEDSS detections remained relatively stable, at approximately 4 to 6 cases. As PADSS case counts increased—particularly more than 150 to 200 cases per week—SEDSS detections rose more sharply, reflecting stronger alignment between systems during peak transmission periods. Data sources: PADSS 14 ; SEDSS.11-13
Cross-correlation analysis showed moderate alignment between weekly SEDSS and PADSS case trends, with the strongest correlations between lag –7 (r = 0.50) and 0 (r = 0.39), suggesting SEDSS trends may lead PADSS by several weeks or align in real time (eFigure 1A in the Supplement). The closely aligned cumulative distribution function curves suggest that, despite week-to-week variability, both systems captured the epidemic curve progressing at a comparable pace over time (eFigure 1B in the Supplement).
Demographic Characteristics
SEDSS dengue patients were younger than PADSS dengue patients (median [IQR] age = 22 [14-36] vs 27 [15-48] y; P < .001) (Table 1). A larger proportion of SEDSS patients than PADSS patients were aged 10 to 19 years (36.2% vs 29.3%; P = .006), while PADSS had a larger proportion of adults aged ≥50 years (23.9% vs 10.5%; P < .001). Sex distribution was similar across systems (males: 55.0%, SEDSS; 53.3%, PADSS). PADSS dengue patients were geographically dispersed across Puerto Rico, whereas most SEDSS patients were in San Juan (63.3%) and Ponce (24.1%) health regions, reflecting sentinel site locations. Recent travel (≤14 d) outside Puerto Rico was uncommon (SEDSS, 5.1%; PADSS, 5.9%). SEDSS patients had earlier specimen collection than PADSS patients for dengue testing relative to symptom onset (median [IQR]: 3 [2-4] vs 4 [2-4] d; P = .003).
Characteristics of confirmed and probable dengue cases (N = 6861) determined by positive test result from RT-PCR, NS1 antigen, or IgM ELISA, by surveillance system, Puerto Rico, January 2024–January 2025 a
Abbreviations: —, not applicable; BMI, body mass index; COPD, chronic obstructive pulmonary disease; DENV, dengue virus; ICU, intensive care unit; IgM ELISA, immunoglobulin M enzyme-linked immunosorbent assay; NS1, nonstructural protein 1; PADSS, Passive Arboviral Disease Surveillance System; RT–PCR, reverse transcription polymerase chain reaction; SEDSS, Sentinel Enhanced Dengue Surveillance System.
The denominator for each variable is the number of participants with available data for that characteristic, which may differ from the total number of participants.
Determined by Pearson χ2 or Fisher exact tests for categorical variables and Mann–Whitney U tests for continuous variables. P values were corrected for multiple comparisons by using the Benjamini–Hochberg procedure; significance set at P < .05.
Data not captured by system.
Denominators for these analyses include visits from women of childbearing age defined as any female aged 15 to 44 years.
Among the 355 SEDSS and 5007 PADSS patients with dengue and available serotype information, DENV-3 was the most common serotype in both systems and was more prevalent in SEDSS (80.8%; n = 286) than in PADSS (59.2%; n = 2965). DENV-2 accounted for 14.6% (n = 52) of SEDSS dengue patients and 12.5% (n = 624) of PADSS dengue patients. DENV-1 was more prevalent in PADSS than in SEDSS (prevalence = 28.3% [n = 1417] vs 4.8% [n = 17]; P < .001), likely reflecting its predominance in western Puerto Rico, where no SEDSS sites are located. A single patient with DENV-4 was reported in PADSS after travel outside Puerto Rico. DENV-3 cases increased in SEDSS in April and May, preceding a sharper rise in PADSS in June (eFigure 2 in the Supplement).
Hospitalization rates were higher in PADSS than in SEDSS (53.1% vs 36.7%; P < .001) (Table 1). Hospitalized SEDSS patients were younger than PADSS patients (median [IQR] age: 16 [13-20] vs 24 [15-48] y; P < .001) (eTables 2-3 in the Supplement). In SEDSS, hospitalization was more frequent among patients with dengue than among patients without dengue (36.7% vs 8.9%; P < .001) (eTable 4 in the Supplement). While PADSS consistently reported more dengue hospitalizations than SEDSS, SEDSS trends generally mirrored the PADSS epidemic curve without providing early warning (eFigure 3 in the Supplement). PADSS hospitalization rates remained high with a nonlinear trend, while SEDSS proportions were more variable given low case counts and showed no clear temporal pattern. Cross-correlation of weekly hospitalized case counts showed modest alignment but no clear lead–lag relationship.
No dengue-related deaths occurred in SEDSS, whereas PADSS recorded 13 deaths among patients with laboratory-confirmed dengue. Acute respiratory virus co-infections were rare in SEDSS (3.1%; n = 8); no co-infection data were available for PADSS (Table 1). Chronic conditions were common among SEDSS patients, with 34.9% having at least 1 comorbidity, including obesity (28.9%), asthma (18.1%), and hypertension (5.4%); comparable data for PADSS were unavailable.
Clinical Characteristics of Dengue Patients
Symptom profiles and laboratory findings highlight differences in fever, rash, thrombocytopenia, and less commonly reported symptoms such as pruritus and dysgeusia (Table 2). A higher proportion of SEDSS dengue patients than PADSS patients had at least 1 dengue warning sign (71.0% vs 58.6%; P < .001). In PADSS, the proportion with warning signs was relatively consistent across age groups, ranging from 47.0% among children aged <10 years to 60.9% among patients aged 10 to 19 years. In contrast, SEDSS captured higher proportions than did PADSS of warning signs among adults: for example, 74.1% in SEDSS versus 52.5% in PADSS among patients aged 30 to 39 years and 66.7% in SEDSS versus 47.7% in PADSS among patients aged ≥50 years (Figure 3). Among hospitalized patients, warning signs were more frequent in SEDSS than in PADSS (86.9% vs 69.3%; P < .001), based on information from the case investigation, convalescent sample processing, and hospital-admitted abstraction forms (eTable 3 in the Supplement). Warning signs more frequently reported in SEDSS than in PADSS were restlessness (31.1% vs 7.5%), mucosal bleeding (21.7% vs 6.9%), and hemoconcentration (4.4% vs 0.1%) (P < .001 for all) (eFigure 4 in the Supplement).
Clinical characteristics of confirmed and probable dengue cases (N = 6861) determined by positive test result from RT-PCR, NS1 antigen, or IgM ELISA, by surveillance system, Puerto Rico, January 2024–January 2025 a
Abbreviations: —, not applicable; IgM ELISA, immunoglobulin M enzyme-linked immunosorbent assay; NS1, nonstructural protein 1; PADSS, Passive Arboviral Disease Surveillance System; RT-PCR, reverse transcription polymerase chain reaction; SEDSS, Sentinel Enhanced Dengue Surveillance System.
The denominator for each variable is the number of participants with available data for that characteristic, excluding unknown responses, which may differ from the total number of participants. Definitions and system-specific differences for these clinical variables are detailed in eTable 1 in the Supplement.
Determined by Pearson χ2 or Fisher exact tests for categorical variables and Mann–Whitney U tests for continuous variables. P values were corrected for multiple comparisons by using the Benjamini–Hochberg procedure; significance set at P < .05.
Data not captured by system.
Severe dengue was defined using World Health Organization 2009 criteria. 16 In PADSS, severe dengue includes any of the following: (1) severe bleeding, (2) severe organ involvement, or (3) physician classification as severe dengue based on clinical judgment. In SEDSS, severe dengue includes any of the following: (1) severe plasma leakage, (2) severe bleeding, or (3) severe organ involvement.
Severe plasma leakage (SEDSS only) was defined as evidence of both respiratory distress and plasma leakage. Respiratory distress included any of the following: tachypnea, labored breathing, accessory muscle use, supplemental oxygen, or intubation. Plasma leakage included any of the following: pleural or pericardial effusion, ascites, free abdominal fluid, hematocrit increase ≥20% during illness or ≥20% above baseline, or low serum albumin for age. Shock was classified as severe plasma leakage if the following were met: documented shock, vasopressor/inotrope/albumin use, or pulse pressure <20 mm Hg or hypotension with ≥2 supportive signs (eg, tachycardia, delayed capillary refill, cyanosis, cold skin). Subcomponents listed under “shock” are elements contributing to the case definition, not stand-alone diagnoses.
Severe organ involvement included encephalitis, coma, or convulsions in PADSS. In SEDSS, it included any of the following: aspartate aminotransferase or alanine aminotransferase ≥1000 IU/L, mechanical ventilation, prothrombin time international normalized ratio ≥1.5, encephalopathy, myocarditis, acute hepatitis, Glasgow Coma Scale <11, aseptic meningitis, or acute paralysis.
Severe bleeding in PADSS was defined as gastrointestinal bleeding or hematemesis. In SEDSS, it included any gastrointestinal bleeding, hematemesis, or receipt of any blood product.
Physician classification as severe dengue: physician-designated severe dengue based on clinical judgment, as indicated on the PADSS case report form.
Percentage of dengue cases, by patient age group in years. (A) Without warning signs. (B) With warning signs without progressing to severe dengue. (C) Severe dengue, by Passive Arboviral Disease Surveillance System (PADSS) and Sentinel Enhanced Dengue Surveillance System (SEDSS), Puerto Rico, January 2024–January 2025. Warning signs were defined according to World Health Organization criteria 18 and included abdominal pain, persistent vomiting, clinical fluid accumulation, mucosal bleeding, restlessness, hepatomegaly, and hemoconcentration. Data sources: PADSS 14 ; SEDSS.11-13
We also observed these trends among hospitalized patients, with higher proportions of warning signs in SEDSS patients than in PADSS patients (eTable 3 in the Supplement). Additional comparisons between SEDSS patients with and without dengue are provided (eTable 5 in the Supplement).
Severe Dengue
The proportion of patients with severe dengue was higher in SEDSS than in PADSS (9.1% vs 5.6%; P = .02) (Table 2), reflecting differences in clinical data collection and definitions. Severe plasma leakage—including cases with respiratory distress or shock—was identified in 6.2% of SEDSS patients but was not captured in PADSS. Among these, 3.5% had documented respiratory distress and 3.5% had clinical evidence of shock; some patients had both features. Severe organ involvement (SEDSS, 1.3%; PADSS, 0.7%) and severe bleeding (SEDSS, 1.9%; PADSS, 3.7%) were uncommon in both systems. In PADSS, we noted a physician-designated classification of severe dengue—based on clinical judgment—for 97 of 659 dengue cases with available data (14.7%). In PADSS, the highest proportion of patients with severe dengue was reported among adults aged 20 to 29 years (7.0%) and 30 to 39 years (6.7%), whereas in SEDSS, the highest proportion was among younger populations: 16.3% among patients aged 10 to 19 years and 10.5% among patients aged <10 years (Figure 3). Shock was identified exclusively in SEDSS. Severe organ involvement, including seizures, encephalitis, and liver impairment, was rare and similarly low in both systems (PADSS, 0.7%; SEDSS, 1.3%).
Discussion
This study highlights the complementary strengths of active and passive surveillance systems in monitoring dengue during the 2024-2025 epidemic in Puerto Rico. PADSS provided broader geographic coverage and higher case counts, while SEDSS offered richer clinical detail and better detection of severe disease. The systems demonstrated moderate temporal alignment, particularly during high-transmission periods when PADSS weekly cases exceeded 100, underscoring the value of sentinel surveillance for tracking epidemic trends and informing timely public health actions such as clinician alerts and targeted vector control. However, SEDSS’s lower sensitivity at lower transmission levels (<100 weekly PADSS cases), when SEDSS detections plateaued at 5 to 6 cases per week, highlights the limitations of relying solely on sentinel sites.
These findings align with prior research using SEDSS to monitor emerging arboviral threats and SARS-CoV-2, where weekly trends in laboratory-confirmed cases were compared across surveillance systems. 20 In the present analysis, the presence of multiple correlation peaks across lags may reflect variability in reporting delays, differences in surveillance processes, or stochastic effects and should be interpreted with caution. Previous work showed that rises in SEDSS case counts preceded those in PADSS by up to 3 weeks during dengue epidemics (2012-2014) and up to 8 weeks during the Zika outbreak, 20 highlighting SEDSS’s potential as an early warning system for rising transmission ahead of island-wide passive reporting. Although this comparison was conducted retrospectively, prior analyses of laboratory turnaround times showed that SEDSS typically returns RT-PCR results several days faster than PADSS does (for dengue, a median of 11 days vs 15 days from specimen collection, and for Zika virus, a median of 4 days vs 14 days), making earlier detection of rising case counts and serotype shifts operationally feasible in real time. 20 Similarly, sentinel systems in settings such as Iquitos, Peru, have successfully tracked dengue trends closely during seasonal surges, reinforcing the value of integrating sentinel and passive approaches for epidemic surveillance. 21
SEDSS captures a broader spectrum of clinical characteristics and severe outcomes than PADSS does, offering insights into disease progression and severity. While collecting detailed clinical data is resource intensive, it helps assess the prevalence of severe dengue, identify populations at risk of severe dengue, and guide resource allocation, interventions, and clinical management.11,22 Detailed clinical data are especially valuable during serotype shifts or the introduction of new DENV serotypes, when clinical presentations may change. In this study, SEDSS detected more severe outcomes and warning signs, especially in younger populations, including severe plasma leakage and shock—both absent in PADSS. Shock was identified in SEDSS by using structured abstraction based on multiple clinical indicators: documented shock, vasopressor or albumin use, pulse pressure <20 mm Hg, and supportive signs such as cyanotic limbs, tachycardia, delayed capillary refill, and mottled or cold skin. Although more consistent than the methods used in PADSS, this approach relied on retrospective medical record review rather than direct physician diagnosis and may not fully reflect clinical judgment. Still, it underscores the value of SEDSS in detecting patients at high risk of severe outcomes and guiding public health decisions on clinician training, planning for surges in pediatric cases, and targeted messaging during outbreaks.
SEDSS’s structured documentation captured some warning signs, such as restlessness, which can be challenging to capture in passive systems such as PADSS. In SEDSS, restlessness included nervousness, agitation, irritability, or disorientation—potential signs of neurologic involvement, hemodynamic instability, or worsening clinical status. In contrast, PADSS grouped these signs under broader categories such as lethargy and fatigue. While lethargy may reflect more advanced neurologic involvement or clinical deterioration, SEDSS’s documentation approach may increase sensitivity to earlier signs of worsening illness, including restlessness, agitation, or disorientation. Recent PADSS updates, through the integration of an electronic case-reporting system, added clinical variables such as shock, bleeding requiring intervention, admission to an intensive care unit, and transfusion history and improved the system’s ability to detect severe dengue. Still, variability in PADSS documentation, application of case definitions in clinical settings, and laboratory testing protocols can contribute to underreporting and inconsistent classification, underscoring the need for standardization. 23 SEDSS’s structured approach helps minimize these discrepancies, enabling the earlier detection of severe cases and timelier public health responses compared with PADSS. Conversely, PADSS may overrepresent severe or hospitalized cases because reporting is often triggered by clinicians when patients are more visibly ill or require hospital-based care, leading to a bias toward capturing more severe cases. SEDSS, by actively enrolling febrile patients regardless of severity, provides a more realistic estimate than PADSS of hospitalization rates among all patients with symptomatic dengue.
SEDSS captured younger dengue patients than PADSS did, which may reflect differences in the populations served by sentinel sites, including emergency departments with substantial pediatric patient volume, and demographic differences in San Juan and Ponce, where SEDSS is based. In contrast, PADSS’s broader geographic coverage, compared with SEDSS, included older populations who may seek care in nonsentinel settings or have different health care–seeking patterns. Higher thrombocytopenia rates in PADSS than in SEDSS may be an artifact of clinician-driven reporting, where platelet counts are more likely recorded when thrombocytopenia is suspected. Documentation practices may also contribute to this discrepancy. The predominance of DENV-3 in SEDSS aligns with recent serotype shifts in Puerto Rico, highlighting the value of sentinel systems in tracking emerging serotype patterns for vaccine and intervention planning. 9 Integrating genomic surveillance into SEDSS could enhance its utility by enabling near–real-time detection of DENV variants. Previous genomic surveillance efforts in Puerto Rico successfully traced the emergence and spread of SARS-CoV-2 and Zika virus,24-26 demonstrating how genomic epidemiology can improve arboviral surveillance. In addition to dengue, both surveillance systems have the capacity to detect other arboviruses, including chikungunya, Zika, and, more recently, Oropouche virus (OROV). SEDSS initiated targeted OROV testing in 2024 for eligible participants in response to regional reports, underscoring its flexibility to adapt testing protocols for emerging threats, although we detected no OROV cases during our study period.
The unusual transmission profile of this dengue event—sustained above-threshold incidence for more than a year—may signal shifts in dengue epidemiology in Puerto Rico. Whether this shift indicates a new normal of persistent high-level transmission or a prolonged prelude to a future peak in 2025-2026 remains uncertain. Continued monitoring will determine whether the trends we observed evolve into a different epidemic pattern, and additional noteworthy findings may yet emerge as the event progresses. For this reason, we interpret our findings as a characterization of system performance during an ongoing high-transmission period, rather than a complete description of the epidemic’s trajectory.
Limitations
This study had several limitations. First, although SEDSS provides detailed clinical data, findings may not be generalizable beyond sentinel sites or to regions with different health care access or surveillance practices. Second, PADSS data may be biased by underreporting or inconsistent documentation of symptoms and outcomes. Third, SEDSS representativeness is limited by its geographic scope and recruitment capacity, especially during high-level transmission periods when sites may be overwhelmed. For example, a drop in SEDSS case detection in late December 2024 likely reflected holiday-related staffing shortages rather than true declines in transmission. Fourth, severity definitions relied on retrospective clinical variables, which may not fully reflect clinical judgment, potentially leading to misclassification. Finally, we did not fully account for spatiotemporal heterogeneity in clinically detected dengue (care seeking, testing, and reporting) or in underlying DENV transmission in the community (asymptomatic and untested infections). Both can vary by place and time and may influence apparent system performance and lead–lag relationships.
Conclusion
Our findings reaffirm the value of integrating sentinel and passive surveillance for epidemic preparedness.21,23,27,28 SEDSS captures early transmission trends and detailed clinical data, whereas PADSS provides broad geographic coverage, together enhancing epidemic response. As arboviral threats grow, Puerto Rico’s integrated approach offers a model for strengthening surveillance of vector-borne diseases. Improving SEDSS’s representativeness and scalability could further enhance its utility. For example, its 2018 expansion to the San Juan Metro area improved coverage and early SARS-CoV-2 detection. These findings highlight the need to balance investments in both systems to optimize public health action. 29
Supplemental Material
sj-docx-1-phr-10.1177_00333549261440198 – Supplemental material for Leveraging Sentinel Enhanced Surveillance to Characterize the 2024-2025 Dengue Epidemic in Puerto Rico: Insights and Comparisons With Passive Surveillance
Supplemental material, sj-docx-1-phr-10.1177_00333549261440198 for Leveraging Sentinel Enhanced Surveillance to Characterize the 2024-2025 Dengue Epidemic in Puerto Rico: Insights and Comparisons With Passive Surveillance by Zachary J. Madewell, Jomil Torres Aponte, Carla Espinet, Dania M. Rodriguez, Olga Lorenzi, Janice Perez-Padilla, Verónica M. Frasqueri-Quintana, Vanessa Rivera-Amill, Diego Sainz de la Peña, Jorge Bertrán Pasarell, Fhallon Ware-Gilmore, Gilberto A. Santiago, Laura E. Adams, Gabriela Paz-Bailey, Melissa Marzan-Rodriguez and Liliana Sánchez-González in Public Health Reports®
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by Centers for Disease Control and Prevention (CDC) grant numbers U01CK000473 and U01CK000580 (V.R.A.).
Disclaimer
The findings and conclusions in this article are those of the authors and do not necessarily represent the official position of CDC.
Supplemental Material
Supplemental material for this article is available online. The authors have provided these supplemental materials to give readers additional information about their work. These materials have not been edited or formatted by Public Health Reports’s scientific editors and, thus, may not conform to the guidelines of the AMA Manual of Style, 11th Edition.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
