Share this article on:

Effects of Patient Medication Requests on Physician Prescribing Behavior: Results of A Factorial Experiment

McKinlay, John B. PhD*,†; Trachtenberg, Felicia PhD*; Marceau, Lisa D. MPH*; Katz, Jeffrey N. MD, MSc‡,§; Fischer, Michael A. MD

doi: 10.1097/MLR.0000000000000096
Original Articles

Background: Because of internet searches, advice from friends, and pharmaceutical advertising, especially direct-to-consumer advertising, patients are increasingly activated to request medications during a physician encounter.

Objectives: To estimate the effect of patient requests for medications on physician-prescribing behavior, unconfounded by patient, physician, and practice-setting factors.

Research Design: Two experiments were conducted among 192 primary care physicians, each using different video-based scenarios: an undiagnosed “patient” with symptoms strongly suggesting sciatica, and a “patient” with already diagnosed chronic knee osteoarthritis. Half of patients with sciatica symptoms requested oxycodone, whereas the other half requested something to help with pain. Similarly, half of knee osteoarthritis patients specifically requested celebrex and half requested something to help with pain.

Subjects: To increase generalizability and ensure sufficient numbers were available, we recruited 192 primary care physicians from 6 US states.

Measures: The primary outcome was whether physicians would accede to a patient’s request for a medication. Alternative pain medications prescribed were secondary outcomes.

Results: 19.8% of sciatica patients requesting oxycodone would receive a prescription for oxycodone, compared with 1% of those making no specific request (P=0.001). Fifty-three percent of knee osteoarthritis patients requesting celebrex would receive it, compared with 24% of patients making no request (P=0.001). Patients requesting oxycodone were more likely to receive a strong narcotic (P=0.001) and less likely to receive a weak narcotic (P=0.01). Patients requesting celebrex were much less likely to receive a nonselective nonsteroidal anti-inflammatory drugs (P=0.008). No patient attributes, physician, or organizational factors influenced a physician’s willingness to accede to a patient’s medication request.

Conclusions: In both scenarios, activated patient requests for a medication substantially affected physician-prescribing decisions, despite the drawbacks of the requested medications.

Supplemental Digital Content is available in the text.

*New England Research Institutes (NERI), Watertown

Division of Medicine, Harvard Medical School, Massachusetts General Hospital

Orthopedic and Arthritis Center for Outcomes Research, Department of Orthopedic Surgery and Division of Rheumatology, Brigham and Women's Hospital, Harvard Medical School

§Department of Epidemiology, Harvard School of Public Health

Division of Pharmacoepidemiology and Pharmacoeconomics, Brigham & Women's Hospital, Harvard Medical School, Boston, MA

Supplemental Digital Content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Website,

Supported by Award Number AR056992 from the National Institute of Arthritis and Musculoskeletal and Skin Disorders (NIAMS) (NIH).

Dr Fischer has received research support from CVS-Caremark for studies of medication adherence. All other authors attest that they have no financial interest conflicting with complete and accurate reporting of the study findings.

Reprints: John B. McKinlay, PhD, Health Services & Disparities Research, New England Research Institutes, 9 Galen Street, Watertown, MA 02472. E-mail:

© 2014 by Lippincott Williams & Wilkins.