Do patient medication requests affect primary care physician prescribing? Results from a factorial experiment

Because of exposure to pharmaceutical advertising in different media outlets, especially direct to consumer advertising (DTCA), patients are increasingly activated to request a specific medication during an encounter with a physician. To estimate the effect of patient requests for specific medications on physician prescribing behavior, unconfounded by selected patient attributes (age, race/ethnicity, SES), physician characteristics (gender, experience) and many practice setting factors. Employing balanced factorial experimental methods two clinically authentic video-based scenarios were presented: an undiagnosed “patient” with symptoms strongly suggesting sciatica, and a “patient” with already diagnosed chronic osteoarthritis of the knee. Half of the patients with sciatic symptoms specifically requested oxycodone, while the other half simply requested something to help with pain. Similarly, half of those with knee osteoarthritis specifically requested Celebrex. Community-based primary care physicians from several different US states were recruited as experimental subjects (n=192). The primary outcome was whether physicians would accede to a patient request for a specific medication. Alternative pain medications that would be prescribed were secondary outcomes. 19.8% of the sciatic patients specifically requesting oxycodone would receive a prescription, compared with just 1% of those making no request (p=0.001). 53% of patients with knee osteoarthritis requesting celebrex would receive it, compared with 24% of patients making no specific request (p=0.001). Assertive 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), or an NSAID. Assertive patients requesting celebrex were much less likely to receive a narcotic (p=0.008), or an NSAID. None of the patient attributes, the physician characteristics or the organizational factors influenced a physician's willingness to accede to a patient request for a specific medication. Patients making a request for a specific medication are highly likely to have it prescribed, with the high likelihood of receiving the requested medication consistent across the two scenarios. Who the patient and physician are, and features of the practice setting in which the encounter occurs, have no effect on the success of a medication request. This study suggests that the impact of activated patients on physician decisions potentially increases the total number of patients who become activated and make specific requests of their physicians. Our findings indicate that these requests are often likely to succeed. Since DTCA is used exclusively for expensive medications, generally those like celecoxib that are still available only in branded forms, this effect is likely to increase medication costs.