A pilot study in Sofia found that sales of selected over-the-counter respiratory products were generally higher in January and broadly overlapped with the city's reported seasonal acute respiratory infection peak. The finding shows that pharmacy sales information can be collected in practice, but it does not establish an association with infection activity or show predictive performance.
The project asked whether community-pharmacy sales could serve as a complementary indicator of respiratory infection activity. Researchers used retrospective monthly totals from two community pharmacies in Sofia, covering October 2024 through April 2025, one respiratory-infection season. The data were analyzed at product-category level, without identifying brands or individual pharmacies.
The product list was built in stages. Researchers reviewed prime-time Bulgarian television advertising during February and March 2025, then sought expert review and pharmacist input before grouping products into respiratory-symptom categories. A separate selection survey included 97 community pharmacists.
January stood out, but not decisively
Antipyretics, a category of fever medicines, made up the largest monitored sales category in both pharmacies. The monthly counts ranged from 22 antiviral sales in April 2025 to 503 antipyretic sales in January 2025. Antipyretics represented 35.74% of included sales in Pharmacy 1 and 74.45% in Pharmacy 2; nasal congestion and runny-nose remedies accounted for 28.29% in Pharmacy 1.
Average monthly sales were much higher for antipyretics, at 685.96 units, than for antivirals, at 121.06. But antivirals had the highest coefficient of variation, a measure of how uneven sales were relative to their average, at 64.78, compared with 22.8 for antipyretics.
January accounted for 19.14% of all antipyretic sales and 32.00% of all antiviral sales recorded across the seven-month period. The January share of antiviral sales was 29.80% in one pharmacy and 35.36% in the other, while supplements made up 19.66% in January, almost identical to 19.65% in October 2024.
An overall Friedman test, which compares repeated monthly distributions, found a statistically significant difference across the months, with January ranked highest and April lowest. That result did not settle whether January was a genuine peak. In follow-up comparisons, January did not differ significantly from any of the other six months after correction for multiple testing; several unadjusted p-values were about 0.012, but the reported adjusted threshold was 0.0083. The authors therefore treated the January increase as descriptive rather than statistically confirmed.
A collection test, not a forecasting system
The analysis summarized monthly category sales using averages and measures of spread, then tested overall month-to-month differences and January against the other months. Because the observations were monthly and limited, the researchers did not run formal time-series, lag or cross-correlation analyses. They therefore did not assess whether sales changes came before routine surveillance signals or estimate predictive performance.
The January sales pattern broadly coincided with the reported Sofia acute respiratory infection peak in week 4 of 2025. That is temporal concordance, meaning that the two patterns appeared in the same general period. It does not show that one caused the other, that pharmacy sales were associated with infection incidence, or that the sales could serve as a validated early-warning signal.
The authors describe the work as an initial feasibility assessment. Its main contribution was showing that pharmacy sales information could be gathered and organized for analysis, not validating an operational early-warning indicator. The paper's own design left accuracy, lead time and surveillance performance untested.
Small sample, wide uncertainty
The evidence base was narrow: only two community pharmacies contributed data during one respiratory season. The authors say those pharmacies were not representative of Sofia or Bulgaria, limiting how far the findings can be generalized.
Several measurement problems also complicate interpretation. The advertising period used to select products did not match the full sales period; categories were not based on a formal medicine-classification system known as ATC; and product availability differed between pharmacies. The researchers also lacked information on stock levels, shortages, prices, promotions, customer characteristics, prescriptions and clinical diagnoses. Monthly aggregation prevented formal time-series, correlation and lag analyses.
For now, the study supports feasibility rather than deployment. Larger prospective work with more representative pharmacies, standardized product categories and more frequent data would be needed to determine whether sales correlate with infection incidence or provide an earlier signal. The current results do not support using these sales as an operational surveillance system.
The authors reported financial support from the Southeast European Influenza Regional & Country Specific work led and coordinated by SECID, PIVI Bulgaria, and the Center of Competence ImmunoPathogen programme.
Paper data and sources
Original title: Feasibility of using community pharmacy sales of over-the-counter respiratory products as a syndromic surveillance indicator: a pilot study from Sofia, Bulgaria
Authors: Getsova Z, Rangelova V, Sapunarov V et al.
Journal/Repository: Frontiers in public health
Status: Peer-reviewed
First online: 2026-10-09
DOI: Not available
Original paper · Full text