Jessi Espino MD MS
The RODS Laboratory has conducted multiple published studies to understand the relationship between medication sales and epidemics.
If the patient decides not to seek treatment, he will seek other approaches or delay treatment. Therefore, data sources such as over-the-counter drug sale or self-treatment merchandise sales may provide the only information about these sick individuals.
– Zeng and Wagner 2001
We started with a review of the literature. In 2001, Dr. Xiaoming Zeng (then a graduate student at the RODS Laboratory) created a model of patient health-seeking behavior. In this model he found that “if the patient decides not to seek treatment, he will seek ther approaches or delay treatment. Therefore, data sources such as over-the-counter drug sale or self-treatment merchandise sales may provide the only information about these sick individuals” This work was published in Zeng X, Wagner M. Modeling the effects of epidemics on routinely collected data. Proc AMIA Symp. 2001:781-5. PMID: 11825292; PMCID: PMC2243441.

Additional work that year by Carnegie Mellon University colleague Dr. Anna Goldenberg used data from a grocery store chain and verified our initial hypothesis regarding the relationship between health-seeking behavior and medication purchases. This work can be found in Goldenberg A, Shmueli G, Caruana RA, Fienberg SE. Early statistical detection of anthrax outbreaks by tracking over-the-counter medication sales. Proc Natl Acad Sci U S A. 2002 Apr 16;99(8):5237-40. doi: 10.1073/pnas.042117499. PMID: 11959973; PMCID: PMC122753.
Sales of cough medication have widely varied patterns: a seasonal effect, with winter sales higher and more chaotic than summer sales, a weekly effect showing higher sales during weekends, peak sales on holidays, and low sales on days when many stores are closed. Goldenberg (2002)

With our initial research complete, we set out to create the National Retail Data Monitor (NRDM). The system was designed to collect over-the-counter medication sales (OTC) data on a large scale for public health. Like our early experiences building the Real-time Outbreak and Disease Surveillance System, we knew that only by building such a system could we verify our hypotheses about the large scale effect of epidemics on medication sales data and determine the limitations of such a data source. In the process we hoped to provide an invaluable resource and service to public health.
We quickly saw that OTC data collected by the NRDM would be inherently useful for influenza and seasonal gastrointestinal outbreak tracking. Along with emergency department surveillance efforts using the RODS system we saw regular patterns in the medications sales data that would be useful for anticipating coming epidemics.
For influenza monitoring, we saw that across multiple states and multiple years OTC medication sales data was highly correlated with visits to emergency rooms for respiratory and constitutional illnesses. Moreover, the OTC sales data was inherently more timely than visits to emergency rooms.

An unanticipated find was the strong effect of pediatric respiratory and gastrointestinal epidemics on the OTC medication sales. In this landmark study, we found that OTC Electrolyte sales preceeded hospital visit for gastrointestinal and respiratory illnesses by 2.4 weeks. This work can be found in Hogan WR, Tsui FC, Ivanov O, Gesteland PH, Grannis S, Overhage JM, Robinson JM, Wagner MM; Indiana-Pennsylvania-Utah Collaboration. Detection of pediatric respiratory and diarrheal outbreaks from sales of over-the-counter electrolyte products. J Am Med Inform Assoc. 2003 Nov-Dec;10(6):555-62. doi: 10.1197/jamia.M1377. Epub 2003 Aug 4. PMID: 12925542; PMCID: PMC264433.
Electrolyte sales preceeded hospital visits for gastrointestinal and respiratory illnesses by 2.4 weeks.
– Hogan (2003)
These multi-jurisdiction, multi-year studies were extremely strong evidence of the value of OTC data for outbreak monitoring. These correlations are so strong and so timely that the data should be used as primary data source for anticipating seasonal outbreaks.
For skeptics, these strong correlations are “not enough” to convince them of the value of such data. My colleague, Heather Johnson performed a study to find out first-hand the relationship between medication sales and health-seeking behavior. In her study she personally interviewed pediatric patients and their caregivers at a Pittsburgh emergency room. She found that of the 78 children in the study, 41% of caregivers purchased OTC medications prior bringing them to the ED. They purchased these medications an average of 1.88 days before their visit.
41% of caregivers purchased OTC medications prior bringing them to the ED. They purchased these medications an average of 1.88 days before their visit.
– Johnson
The National Retail Data Monitor continues to be an invaluable resource to public health. We encourage public health to use the system as a resource for public health practice, planning and research.