Expert Advice: Understanding Pharmacist Recommendations for branded versus Generic Over-the-Counter medication

The study varies how a consumer requests advice from a pharmacist about an over-the-counter pain medication. Participating pharmacy locations are exposed to two standardized consumer inquiry conditions on separate visits. One condition allows the pharmacist to provide an unsolicited product recommendation, while the other makes the choice between alternative versions of the medication more explicit. The order of the two conditions is varied across pharmacy locations. This design allows us to examine whether the framing of a consumer’s request affects the advice provided by pharmacists.

Each pharmacy is visited twice and receives both experimental conditions, with the order of conditions randomly assigned across pharmacy locations.

Open-ended condition: The researcher poses as a consumer and asks the pharmacist: “My sister has a toothache. What would you recommend for the pain?” No product is initially presented. This condition is designed to capture the pharmacist’s default recommendation, including whether the pharmacist recommends a branded product, a generic medication/active ingredient, both, or another option.

Brand-versus-generic condition: The researcher presents a branded and generic version of the same OTC pain medication (e.g., Advil and generic ibuprofen) and asks: “My sister has a toothache. Which one of these would you recommend she takes?” Where appropriate, the interaction includes the standardized follow-up: “Is Advil more expensive because it’s better?”

The second visit uses the condition not administered during the first visit. Visits are separated by approximately four to six weeks and, where feasible, a different researcher conducts the second visit to reduce recognition and recall. Researchers follow standardized interaction protocols. The paired design generates within-pharmacy variation in consumer framing, allowing recommendations under the two conditions to be compared while holding relatively stable pharmacy-level characteristics constant.

2025-11-14

2026-08-31

Pharmacist recommendation of branded versus generic OTC pain medication. The primary outcome is the type of medication recommended by the pharmacist during the interaction, classified as a generic medication/active ingredient, a brand-name product, both/either, or no clear recommendation.

The pharmacist’s recommendation will be coded from the recorded interaction and structured data collection form. The principal outcome for the quantitative analysis will be Generic Recommendation, coded as 1 when the pharmacist recommends a generic OTC medication and 0 when the pharmacist recommends a branded medication. Responses in which both/either is recommended or no clear recommendation is provided will be identified separately and summarized descriptively. The primary analysis will compare recommendation behaviour between the Open-Ended and Brand-versus-Generic conditions using the paired, within-pharmacy study design.

Content and reasoning accompanying the pharmacist’s recommendation. Secondary outcomes include whether the pharmacist mentions brand-generic equivalence, cost, effectiveness, brand familiarity, product availability, the active ingredient/ibuprofen, or product location, as well as any other reasoning offered. We will also examine physical indications of product endorsement and responses to the standardized follow-up question in the Brand-versus-Generic condition.

Secondary outcomes will be constructed from the structured coding of each pharmacist interaction. Separate indicators will identify whether the pharmacist mentions equivalence, cost, effectiveness, brand familiarity, availability, active ingredient/ibuprofen, or product location. Additional information on other reasoning, physical indication of a product, and the response to the follow-up question will also be recorded and analyzed where applicable. These outcomes will be compared across the two consumer inquiry conditions to examine whether question framing affects not only which product pharmacists recommend, but also the information and reasoning they provide.

This study uses a naturalistic field experiment in community pharmacies in Ottawa, Canada. Each pharmacy is observed under two different consumer inquiry conditions on separate visits. The order of the conditions varies across pharmacy locations. This paired design allows us to compare how the same pharmacy responds to different ways of framing a request for advice while reducing the influence of stable differences across pharmacy locations. The study focuses on whether consumer framing affects pharmacists’ recommendations between branded and generic over-the-counter pain medications and the information they provide alongside those recommendations.

The study uses a within-pharmacy paired audit design. Each pharmacy is visited twice and receives both experimental conditions: an Open-Ended inquiry and a Brand-versus-Generic inquiry. The order of the two conditions is randomly assigned across pharmacy locations, so that some pharmacies receive the Open-Ended condition first and others receive the Brand-versus-Generic condition first. The second visit uses the alternate condition.

In the Open-Ended condition, the researcher poses as a consumer and asks what the pharmacist would recommend for a sister with a toothache, without initially presenting a specific medication. In the Brand-versus-Generic condition, the researcher presents branded and generic versions of ibuprofen and asks which product the pharmacist would recommend. A standardized follow-up question about the price difference may also be asked.

Repeat visits are separated by approximately four to six weeks. Where feasible, a different researcher or research assistant conducts the second visit to reduce recognition and recall. Researchers follow standardized scripts and interaction protocols and do not disclose the research purpose during the interaction.

The paired design allows within-pharmacy comparison of recommendation behavior across consumer framings. The study records the pharmacist’s recommendation as well as contextual characteristics including pharmacy type, owner status, queue length, number of staff, time of day, weather conditions, researcher identity, and other interaction-level information. The analysis will compare recommendation outcomes across conditions and may adjust for these observed contextual characteristics.

Pharmacies were selected from a master list using stratified random selection based on geographic location and pharmacy type. Selected pharmacies were then randomly assigned to receive either the Open-Ended condition or the Brand-versus-Generic condition on the first visit. Each pharmacy receives the alternate condition on the second visit.

Pharmacy location (site). Randomization was conducted at the pharmacy-site level. The randomized element is the order in which each pharmacy receives the two consumer inquiry conditions.

No

100 pharmacy locations, with two interactions per pharmacy, for a planned total of approximately 200 pharmacist interactions.

200

100 BrandvsGeneric
100 Open-ended

At a 5% significance level and 80% statistical power, with a planned sample of approximately 100 pharmacy locations observed under both consumer inquiry conditions, the study is designed to detect a minimum difference of approximately 15–20 percentage points in brand-versus-generic recommendation rates between conditions. The exact minimum detectable effect depends on the proportion of pharmacies whose recommendations differ between the two conditions, reflecting the paired within-pharmacy design.