Tracking animal bites to forecast rabies treatment demand
Animal-bite injuries are a regular presentation in Australian emergency departments, especially after dog interactions involving children in parks, households and other public places. The health concern is not every bite: clinicians must decide whether an exposure warrants wound treatment, rabies vaccine, human rabies immunoglobulin (HRIG), or simply observation and advice. Surveillance helps health authorities estimate that treatment workload before individual diagnoses or laboratory results arrive.
Australia has no endemic terrestrial rabies, but past imported cases, bat-related exposures and clusters involving a single animal show why preparedness cannot be assumed. Monitoring emergency department visits for animal bites provides early warning of unusual exposure patterns and can reveal pressure on vaccine and HRIG stocks. A rise in presentations may also reflect a local increase in dog or cat behaviour, an animal-control issue or better reporting rather than a genuine surge in rabies risk.
This approach supports population-level planning rather than diagnosing rabies in the emergency room. It can show where and when exposure assessments are increasing, identify settings needing investigation, and give public health teams time to align clinical advice, laboratory testing and supply. The central measure is demand for rabies post-exposure prophylaxis, interpreted alongside species, bite location, vaccination history and the likelihood of virus exposure.
Define a syndromic animal-bite signal
A useful syndrome begins with a clinical complaint rather than a positive rabies test. Include ED encounters where the presenting problem is an animal bite or scratch, then stratify by species when available: dog, cat, bat, fox or another mammal. Exclude codes clearly related to insect bites, non-animal trauma and repeat wound checks when the aim is to count new exposures. A separate flag should capture whether the animal can be identified, observed and tested.
Record a consistent minimum dataset covering age, postcode or statistical area, triage category, species, exposure setting, wound location and severity, time of injury, prior rabies vaccination and treatment provided. Free-text clinical notes can add important details, but they should be processed through structured extraction and privacy controls. Exact addresses are unnecessary for population surveillance, while broad location data can support investigation of clusters without identifying individual patients.
The definition should separate three questions: how many people presented with an animal injury, how many exposures had features that could justify rabies assessment, and how many people actually began PEP. Most dog and cat bites in rabies-free settings do not require rabies prophylaxis, while direct bat contact or an animal with unusual behaviour needs specialist public-health assessment. Confusing these categories would produce a misleading estimate of treatment demand.
Turn emergency visits into a demand forecast
Extract and clean data from triage diagnoses, discharge codes, presenting complaints and relevant free-text fields. Emergency departments differ in coding and documentation, so a national data dictionary should map equivalent terms while retaining local detail. Methods used for respiratory outbreak detection demonstrate how longitudinal ED data can reveal changes that are difficult to recognise from laboratory-confirmed cases alone.
Measure weekly counts and rates rather than relying on raw totals alone. A useful dashboard compares each area with its historical baseline, a rolling four-week period and the capacity of nearby treatment facilities. Stratification by species, age group, exposure setting and urgency can reveal a clinically meaningful cluster hidden in an otherwise modest rise in attendances. Data-quality notes should accompany every release because changes in coding or reporting coverage can resemble a true increase.
A practical forecast has two layers: observed demand for animal-bite assessment and estimated demand for rabies-specific treatment. Historical clinical audits can estimate the proportion of presentations likely to require vaccine or HRIG, after which the forecast can adjust for local circumstances such as a cluster involving one aggressive animal. Models should express uncertainty and never infer rabies risk from attendance counts alone. Stock planning should also distinguish first doses from multi-dose vaccine courses and include contingency for people presenting late.
Interpret signals in Australian settings
Interpretation must reflect local habits and exposure patterns. Sydney and Melbourne have large dog-owning communities and busy public parks, where after-school activities, dog walking and contact between unfamiliar dogs can create clusters of bite injuries. Brisbane and other Queensland communities have flying-fox colonies, making careful assessment of bat exposure important even though the presence of bats does not demonstrate rabies. In rural and remote parts of Western Australia, Tasmania and northern Australia, distance from specialist services and delayed presentation can increase both the need for clear referral pathways and the importance of holding treatment supplies locally.
Context signals that strengthen interpretation
- Weekly ED bite visits per 100,000 residents
- Species, wound site and exposure setting
- Repeated links to one animal, household or public place
- Concurrent animal-control calls and ambulance dispatches
Legal and supply arrangements also shape what a signal means. Rabies is notifiable across Australian states and territories, while local councils apply animal-management rules that vary by location. The domestic market for rabies vaccine and HRIG is small and usually operates through public health units or designated hospitals rather than routine retail pharmacy shelves. A modest rise in orders can therefore be operationally important even when ED attendance remains low.
Measures of treatment demand
- Clinician-coded exposures assessed as eligible for PEP
- HRIG requests by hospital and public health unit
- First vaccine doses and completion of multi-dose courses
- Short-term stock use in areas linked to a cluster
Do not convert every animal-bite encounter into a forecast of prophylaxis. Free-text extraction may mistake an insect bite for an animal exposure, while changes in triage terminology can create artificial surges. Signal verification should therefore involve clinicians, animal-control officers and public-health teams familiar with the affected area. Exposure patterns also need review after unusual events, including repeated bites at a school, transport hub, camping site or other location where a particular animal may have been present.
Connect emergency data with other surveillance channels
Emergency data should sit within a wider surveillance network. Ambulance dispatches, animal-control call centres, general practices, laboratories, community pharmacies, schools and childcare services can each contribute context. For sparsely populated communities, rural e-pharmacy trends illustrate the potential value of indirect data when one hospital or laboratory may be the only reporting point, although medicine sales are not a direct measure of rabies vaccine or HRIG use.
A multi-channel Australian design could connect ED alerts with ambulance call-outs and pharmacy requests for wound-care products, while schools report clusters without unnecessarily identifying children. Community pharmacists can also flag repeated enquiries about bites or travel-related vaccination, providing an early indication that local awareness or activity has changed. During major international events in Sydney, Melbourne and other destination cities, temporary visitor surges may increase multilingual advice needs and after-hours demand, supporting event-specific monitoring without treating every additional attendance as an exposure signal.
Governance must keep the system useful and trusted. Public dashboards should show aggregated data, while secure case-level information is available only to authorised teams conducting follow-up. Information-sharing agreements should define responsibilities, retention periods and escalation routes, and automated alerts should be checked by a trained analyst before they trigger operational action. Data linkage should improve context rather than replace clinical assessment, with every alert evaluated against exposure details, animal behaviour and current public-health advice.
Move quickly from signal to safe response
Turn each signal into a graded response. Routine variation should be reviewed through normal surveillance processes, while a sustained increase above an agreed threshold should prompt data validation and checks with nearby facilities. A sharp rise linked to one animal, household or public setting should trigger immediate coordination with animal control, the affected ED and the relevant public health unit. HRIG stock, vaccine availability and referral arrangements should be checked before the number of presentations begins to rise.
A practical response pathway should specify who receives the alert, how quickly it is acknowledged and what information must accompany it. Clinical decisions should continue to follow current state and territory guidance and the Australian Immunisation Handbook, including appropriate wound care, observation where relevant and specialist advice for bat exposures. Operational teams should distinguish public reassurance from individual clinical advice and ensure that suspected animal cases can be captured, tested and reported through the correct channels.
Build the pathway now by creating a shared reporting template, maintaining a ready stock and communication protocol, and rehearsing the response with emergency departments, animal control, laboratories, pharmacies and public health teams before the next exposure cluster arrives.