Using pharmacy data to spot a COVID-19 resurgence early
A rise in rapid antigen test kit sales can provide an early signal that COVID-19 transmission is increasing in the community. Pharmacies often see people seeking tests before a surge appears in hospital admissions or official case counts, making retail activity a useful component of syndromic surveillance. The signal is indirect, but it can reveal changes in respiratory illness, household concern and demand for diagnostic testing.
For Australia, pharmacy surveillance is especially relevant because people can buy RATs from community pharmacies, supermarket pharmacies, independent chemists and online retailers. Sales may increase when residents hear about outbreaks, prepare for travel, or test before visiting an older relative. A robust monitoring system therefore combines pharmacy data with clinical reports, ambulance activity, school absenteeism, laboratory volumes and wastewater findings.
Why rapid test sales can act as an early signal
Pharmacy purchases reflect behaviour linked to illness, exposure and perceived risk. Someone with a sore throat, fever or cough may buy a test before contacting a GP. A household may purchase several kits after one member develops symptoms, while a residential aged-care worker may test before a shift. When many pharmacies report an unusual increase in demand over a short period, the pattern may indicate growing respiratory activity.
This approach is a form of syndromic surveillance because it tracks a proxy for illness rather than waiting for confirmed diagnoses. Sales data can be available daily or weekly, depending on the retailer and reporting arrangement. That speed can help public-health teams investigate a possible resurgence while the first wave is still developing.
A sales increase does not prove that COVID-19 transmission is rising. People may stock up because kits are discounted, a new workplace policy has been announced, or a large event is approaching. Data analysts must therefore compare current activity with historical baselines, seasonal patterns and nearby indicators. A sudden jump across several locations is more informative than a short-lived increase at one shop.
How an Australian pharmacy signal is built
A useful monitoring programme collects the number of RAT units sold, the number of transactions, the product type and the location of each sale. It should distinguish individual kits from multipacks, because a small number of bulk purchases can distort the apparent number of people testing. Where possible, analysts should examine daily totals, seven-day averages and changes from the same period in previous years.
Australia’s pharmacy market is varied. Large chains, supermarket pharmacies and independent chemists serve different communities, while online sellers may attract customers seeking bulk packs or lower prices. Chemist Warehouse and other high-volume retailers can contribute substantial data, but relying on one brand could miss activity in local pharmacies across regional Queensland, western Sydney or outer Melbourne. A representative system should include multiple channels and account for differences in customer demographics.
The Therapeutic Goods Administration regulates COVID-19 self-tests, and product availability can change as suppliers adjust their ranges. A shortage, product recall or sudden price reduction may affect sales without reflecting infection levels. Monitoring teams should record stock-outs, promotions, pack sizes and changes in the approved product market so that commercial effects are not mistaken for epidemiological signals.
Connecting pharmacy activity with other indicators
The strongest interpretation comes from combining several weak signals. If RAT sales rise alongside increased respiratory presentations at urgent-care clinics, more ambulance call-outs for breathing difficulties and greater laboratory testing volume, the evidence for a COVID-19 resurgence becomes stronger. Analysts can use laboratory testing volume as a complementary measure, particularly when testing practices change over time.
Pharmacy data can also be compared with school absenteeism, aged-care facility reports and workplace illness notifications. A rise in absences across schools in Brisbane or Adelaide may indicate broad respiratory circulation, although it will not identify COVID-19 by itself. Reports from general practices and emergency departments can help show whether the increase is causing more severe illness or mainly reflecting mild infections managed at home.
Wastewater surveillance provides another useful comparison because it measures viral material shed by a population rather than individual testing behaviour. The timing will not always match retail purchases. People may buy tests immediately after media coverage, while wastewater changes may lag or lead depending on sampling and laboratory processing. Differences between these sources are valuable prompts for investigation rather than reasons to discard the data.
Making sense of geography, timing and behaviour
Location matters because pharmacy sales are shaped by local population size, age structure, travel and access to healthcare. A rise in kit purchases around Sydney’s CBD may reflect commuters, visitors or a major event. A similar increase in a remote Northern Territory community may have a different significance because supply deliveries are less frequent and clinical services are farther away.
Timing also needs careful treatment. Demand may climb before Christmas, during winter respiratory season or ahead of school holidays, when families travel between Melbourne, the Gold Coast and regional towns. Sales can fall temporarily when pharmacies close for public holidays, then appear unusually high when they reopen. Daily figures should therefore be smoothed and interpreted against trading hours, public holidays and distribution schedules.
Public messaging influences behaviour in distinctly human ways. A widely shared news report about a new variant can send people to the chemist that afternoon, even when local transmission is unchanged. Australians may describe symptoms casually as feeling “crook” or having a “head cold”, and many will choose a RAT before making a GP appointment. These patterns matter because surveillance measures both disease activity and the public’s response to perceived risk.
Protecting data quality and privacy
Retail surveillance should use aggregated information whenever possible. Public-health teams generally need counts by date, broad area and product category, not names, addresses or payment details. Data-sharing agreements should define how information is transferred, who can access it, how long it is retained and how results are reported. Small-area data may need suppression so that individuals or small facilities cannot be identified.
Quality checks should look for duplicate transactions, unusual bulk orders, missing outlets and abrupt changes in reporting. A pharmacy that changes its point-of-sale system may appear to experience a sudden surge when the difference is technical. Analysts should flag new stores, temporary closures and changes in online fulfilment before drawing conclusions from the trend.
The system should also distinguish a signal from an alert. A signal means that activity differs from the expected pattern and deserves review. An alert may be issued only after epidemiologists assess supporting evidence, local context and the likelihood of bias. This staged process helps avoid unnecessary public anxiety while allowing health authorities to move quickly when several indicators align.
Turning an early warning into action
When pharmacy purchases rise consistently across several Australian regions, public-health authorities can increase diagnostic and clinical awareness without waiting for hospital pressure to build. They may ask sentinel clinics to review respiratory presentations, encourage laboratories to monitor positive results and remind people with symptoms to stay home and test. Health messaging can be targeted to practical settings such as workplaces, schools, aged-care homes and public transport.
The response should remain proportionate to the evidence. A short increase in RAT sales during a promotion does not justify broad restrictions, while a sustained rise accompanied by positive laboratory results may support stronger advice for high-risk groups. Clear communication should explain what is known, what remains uncertain and where people can obtain reliable testing and treatment information.
The same analytical framework can detect other community health patterns. Pharmacy purchases may complement reports of conjunctivitis, and eye infection signals can help identify clusters that would otherwise be noticed only after people seek medical care. University clinics, schools and aged-care services can add further context when unusual symptoms spread through defined populations.
Enhanced surveillance is particularly valuable around major international events, concerts, sporting tournaments and holiday travel periods. Visitors may introduce or acquire infections in crowded settings, while residents may change their testing behaviour before and after an event. Comparing pharmacy data with venue location, transport patterns and clinical reports can help identify emerging clusters without treating every increase in retail demand as a confirmed outbreak.
From retail patterns to public-health intelligence
A mature system treats pharmacy sales as one stream in a multi-channel surveillance network. Its value comes from speed, coverage and the ability to capture people who never attend a clinic. Its limits come from consumer behaviour, unequal access, changing product prices and uncertainty about whether a purchased kit is actually used or reported.
Interpretation can improve when analysts calibrate the data against positive RAT reports, where available, and laboratory-confirmed results. They can also examine the ratio between kit sales and other indicators. For example, rising sales with stable emergency presentations may suggest mild community transmission, while rising sales followed by increased hospital admissions could indicate a more consequential wave.
Methods developed for other syndromic signals can offer useful lessons. Reports from university health centres, including university rash reports, show how symptom patterns and defined populations can support faster investigation before laboratory confirmation is complete. For COVID-19, the equivalent may be a combination of RAT demand, respiratory symptoms, school absences and laboratory positivity.
The practical goal is timely action, not perfect prediction. A pharmacy signal can prompt further testing, improve preparedness and focus communication on communities showing unusual activity. Used with transparent methods and local knowledge, it can help Australia recognise a resurgence earlier and respond before preventable transmission places greater pressure on hospitals and vulnerable people.
Public-health agencies, pharmacy groups, retailers and researchers can strengthen this approach by establishing consistent reporting standards for RAT sales and related respiratory indicators. Share aggregated trends, document changes in stock and pricing, and connect retail observations with clinical and laboratory evidence. Building these links now will make the next unusual increase easier to detect and faster to investigate.