Dengue early warning through hospital admissions
Dengue is a mosquito-borne viral illness that can progress from fever and rash to plasma leakage, bleeding or shock. Australia’s risk is shaped by travel-associated introduction and local transmission where Aedes mosquitoes are established. Far North Queensland, including Cairns and Townsville, has recurrent seasonal dengue activity, while Brisbane, Sydney, Perth and Darwin experience different patterns of imported infection and occasional local spread. Hospital records may reveal clinical patterns before laboratory confirmation identifies the virus.
Syndromic surveillance for dengue using hospital admission for thrombocytopenia and abdominal pain provides an early-warning layer rather than a diagnosis. The pattern is clinically plausible: dengue can lower platelet counts, and abdominal pain may accompany severe disease. Admissions can expose a rise in suspected dengue or a cluster that would otherwise emerge through conventional notification, provided the signal is compared with seasonal and historical baselines.
Japan’s multi-channel system shows the value of combining hospitals with clinics, ambulances, pharmacies, schools, aged care and laboratory reporting. The same principle can strengthen Australian dengue monitoring, but local pathways, coding systems and response structures must guide implementation. A useful hospital signal should trigger faster testing, case review, vector-control assessment and targeted communication while laboratory-confirmed surveillance continues.
Why hospital admissions reveal dengue trends
Clinical progression creates a useful proxy for dengue activity. Falling platelet counts may appear with dengue, while severe abdominal pain can indicate plasma leakage or haemorrhagic dengue. Other causes of thrombocytopenia include malaria, other viral illnesses, immune thrombocytopenia and medication effects. Abdominal pain also occurs with gastroenteritis, appendicitis, kidney stones and pregnancy-related conditions. The indicators therefore need clinical context rather than automatic diagnostic labelling.
Emergency department presentations and inpatient admissions should be analysed together. Some patients with dengue are assessed and discharged, whereas others require admission because of dehydration, bleeding, persistent vomiting or other warning signs. A rise in emergency department attendances may therefore become visible before an admission increase. Conversely, admissions are a biased and delayed measure of infection because they mainly capture people with more severe illness.
Australian public hospital data can support this work through several routes: coded principal diagnoses, other diagnoses, blood test results, triage categories and admission dates. Direct, near-real-time feeds are more useful for outbreak response than retrospective national hospital statistics. Discharge coding alone can miss the clinical timing of illness or reflect a diagnosis made near separation, so epidemiologists should work with clinical coders to map current ICD-10-AM classifications for thrombocytopenia and abdominal pain.
Building a reliable clinical signal
A practical algorithm begins with defining the observation period, geography and population. Plausible cases can include encounters containing a thrombocytopenia indicator, an abdominal-pain indicator or both. Requiring both conditions in the same episode of care may improve specificity but miss cases where outpatient testing, coding or discharge summaries are incomplete. Monitoring each indicator separately, alongside a combined measure, offers a more balanced approach.
The observed count should then be compared with an expected count based on previous wet seasons, the same period in earlier years and local emergency department activity. Statistical process-control methods, such as cumulative sum charts or the Epidemiologic Alarm System, can identify sustained deviations without treating every small fluctuation as an outbreak. Age group, postcode, admission source and day of week help distinguish a broad rise from a cluster concentrated in one location.
Clinical review remains essential when an alert is triggered. Australian hospitals commonly report a full blood count, platelet count and haematocrit, making serial results valuable for identifying patterns compatible with dengue. Where available, early nucleic-acid testing and appropriately timed serology can confirm infection. Negative initial serology does not necessarily exclude dengue, so collection dates, prior infection and vaccination history must be considered.
Joining clinical and community data
Hospital admissions should form one layer of a broader surveillance system. General practitioner records, pharmacy sales, ambulance dispatch data, laboratory notifications, school absences and mosquito surveillance may reveal different stages of the same outbreak. A state surveillance team can compare hospital trends with school absenteeism data, after adjusting for school holidays, seasonal illness and reporting delays. These community signals cannot diagnose dengue, but they can help indicate whether wider transmission may be occurring.
Japan’s experience with event-based monitoring is relevant to Australia because large international gatherings can increase travel-associated introductions. Temporary enhancements around major events can combine travel information, clinical alerts and vector surveillance. Local transmission still depends on mosquito abundance and suitable weather, so imported case numbers should not be interpreted as evidence that local transmission is already established.
Privacy and governance are particularly important when small geographic areas are monitored. Australia’s privacy framework, state health-records legislation and hospital agreements should guide linkage, access and disclosure. Public dashboards can suppress small cells, display rates rather than raw counts and avoid identifying people with rare conditions. Clear rules are also needed for who can view line-list data, initiate testing or order an entomological response.
Learning from school absenteeism patterns
A detailed example of daycare absenteeism patterns shows how non-laboratory information can support early outbreak detection. Although varicella differs from dengue, the surveillance principle is transferable: a sharp departure from expected attendance can justify investigation before every infection is laboratory confirmed. The same approach can be applied to clusters of dengue-like illness, provided the measure is interpreted cautiously.
School and childcare data have important limitations. Families may keep sick children home from respiratory illness, gastrointestinal outbreaks or ordinary holidays. Attendance systems may be delayed, and schools across Australia use different term dates and reporting practices. Public holidays, tropical wet seasons and local school closures must be incorporated into baseline calculations. The strongest signal is a geographically concentrated increase accompanied by hospital, pharmacy or laboratory evidence.
Dengue surveillance should focus on communities where the clinical pattern and mosquito risk overlap. In Cairns or Townsville, a modest rise in thrombocytopenia admissions during the wet season may be more concerning than a larger rise in a city without established dengue transmission. Imported infection in Brisbane or Sydney is operationally different from local transmission in Far North Queensland, even when the symptoms are identical.
A practical Australian operating framework
Hospital surveillance should follow a short, repeatable workflow: receive de-identified data, assign the encounter to a period and area, compare it with an expected baseline, apply clinical criteria and route a verified alert to the response team. The use of paediatric emergency department data for respiratory syncytial virus provides a useful comparator for structuring emergency department indicators, adjusting for seasonal variation and measuring time to action.
A coordinated Australian program should build on several practical measures:
- Establish a governed data pathway linking public hospital emergency departments and inpatient feeds with state and territory surveillance teams, while assigning clear responsibilities for coding, validation and alert review.
- Use tiered alerts based on sustained excess admissions, geographic concentration and clinical severity, with thresholds adapted for tropical Queensland, northern Australia and southern cities rather than applied as one national rule.
- Pair every syndromic alert with confirmatory testing, clinician review and a vector-response assessment so that detection leads to investigation rather than simply increasing dashboard activity.
- Audit performance using time from symptom onset to presentation, time from alert to case review, laboratory confirmation rates, false-alert frequency and the speed of public-health action.
The indicator should be tested against known dengue seasons and periods without unusual transmission. Sensitivity matters because delayed detection can allow infected people to remain infectious while mosquitoes continue biting. Specificity also matters because repeated false alarms can weaken confidence and divert scarce public-health resources. Regular review by clinicians, epidemiologists, entomologists and data specialists keeps the algorithm aligned with changing case definitions, laboratory methods and coding practice.
Turning alerts into coordinated action
An early-warning threshold should have a predefined response pathway. A lower-level alert can prompt hospital clinicians to review recent cases, request timely testing and report relevant travel or exposure information. A higher-level alert can bring together the state health team, laboratories, hospitals, mosquito-control officers and affected primary-care services. Confirmed cases should also inform testing advice, clinical awareness and local vector-control measures, with public communication focused on symptoms and prevention rather than unconfirmed alarm.
The immediate next step is for health departments, hospitals and surveillance partners in dengue-receptive areas to map their existing admission, laboratory and vector data against a shared escalation plan. That partnership can turn a rise in thrombocytopenia and abdominal-pain admissions into faster investigation, coordinated containment and better protection of Australian communities.