Influence of direct to consumer pharmaceutical advertising and patients' requests on prescribing decisions: two site cross sectional survey
(Published 02 February 2002)
Cite this as: BMJ 2002;324:278
- Barbara Mintzes, graduate researcher ()a,
- Morris L Barer, professora,
- Richard L Kravitz, professor and directorc,
- Arminée Kazanjian, acting directora,
- Ken Bassett, senior medical consultanta,
- Joel Lexchin, associate professord,
- Robert G Evans, professora,
- Richard Pan, executive directore,
- Stephen A Marion, associate professorb
- a Centre for Health Services and Policy Research, University of British Columbia, Vancouver BC, Canada V6T 1Z3,
- b Department of Health Care and Epidemiology, University of British Columbia, Canada,
- c Center for Health Services Research in Primary Care, University of California, Davis, Sacramento CA, 95817 USA,
- d School of Health Policy and Management, York University, Toronto, Ontario, Canada M3J 1P3,
- e PC-AWARE, Center for Health Services Research in Primary Care, University of California Davis Medical Centre, Davis
- Correspondence to: B Mintzes
- Accepted 14 January 2002
Only the United States and New Zealand allow advertising of prescription drugs directed at patients. US spending on such advertising grew rapidly during the 1990s, reaching $2.47bn (£1650m) in 2000.1 The dramatic increase in investment by the US pharmaceutical industry is evidence of an expected effect on sales. On the rationale that such advertising provides important information to consumers and patients who may benefit from advertised products, pharmaceutical manufacturers have campaigned in the European Union2 and Canada3 for the relaxing of current regulatory restrictions. We examined the relation between direct to consumer advertising and patients' requests for prescriptions and the relation between patients' requests and prescribing decisions.
Participants, methods, and results
We carried out a cross sectional survey of a cluster sample of primary care patients in Sacramento, California, from March to June 2001 and in Vancouver, British Columbia, from June to August 2000. We used questionnaires to determine the frequency of patients' requests for prescriptions and of prescriptions resulting from requests. Seventy eight physicians participated in the study, 40 in Vancouver (all family physicians) and 38 in Sacramento (14 general internists and 24 family physicians).
Patients were all 18 years and over, spoke English, and provided informed consent. The unit of analysis was a matched set of patient-physician questionnaires covering a single consultation. We estimated adjusted odds ratios using a generalised estimation equation. We classified drugs as advertised to consumers if they were among the 50 drugs with the highest US advertising budgets4 or were described as advertised to consumers in Canadian media reports5 in 1999-2000, or both.
Sixty one per cent of patients attending physicians' offices on preset study days participated (1431 total; 683 in Sacramento and 748 in Vancouver). Patients in the two cities had similar demographic characteristics, socioeconomic status, and attitudes toward the doctor-patient relationship. In both settings, income was higher than average, and 80% were of European descent.
Patients requested prescriptions in 12% of surveyed visits. Of these requests, 42% were for products advertised to consumers. The table provides details of factors associated with requests. Physicians prescribed the requested drugs to 9% (128) of patients and requested advertised drugs to 4% (55) of patients. The prescribing rate was similar for advertised and non-advertised drugs (about 74%).
After we controlled for health status, demographics, socioeconomic status, drug payment, and physicians' sex, specialty, and years of practice we examined the influence of requests on the probability that a patient received a new prescription. Patients who requested a prescription (for advertised and non-advertised drugs) were more likely to receive one (139/175 v 329/1256, odds ratio 8.7, 95% confidence interval 5.4 to 14.2).
We asked physicians: “If you were treating another similar patient with the same condition, would you prescribe this drug?” An answer of “very likely” indicated confidence in choice and “possibly” or “unlikely” indicated some degree of ambivalence. Physicians were ambivalent about the choice of treatment in around 40% of cases when patients requested drugs (advertised and non-advertised, 62/143 v 62/500, 5.4, 3.5 to 8.5) and about half the cases when patients had requested advertised drugs (30/60 v 62/500, 7.1, 4.0 to 12.6) compared with 12% for drugs not requested by patients.
Patients' requests for medicines are a powerful driver of prescribing decisions. In most cases physicians prescribed requested medicines but were often ambivalent about the choice of treatment. If physicians prescribe requested drugs despite personal reservations, sales may increase but appropriateness of prescribing may suffer. Concerns about the value of opening up the regulatory environment to permit direct to consumer advertising in the EU and Canada seem well justified.
We thank Robert Woollard for assistance with recruitment of physicians in Vancouver and the physician questionnaire; the research assistants who administered the survey: Amit Ahuja, Danielle Lapointe, Michael Tsang, Christine Choi, Vanphen Chanthalangsy, Min H Ku, Laura Shively, Erica Stranger, Nicollet Knopf, Bryan Faulstich, Karry Nagai, and Meridith Cobari; Sara Lu Vorhes and Valerie Olson for research coordination in Sacramento; and Nhue L Do and Alicia Mintzes for data entry.
Members of a multidisciplinary expert advisory panel assisted with the study design: Wendy Armstrong, Alan Cassels, Jean-Pierre Gregoire, Matthew Hollon, Patricia Kaufert, Joel Lexchin, Bob Nakagawa, Nancy Ostrove, Richard Pollay, and Ingrid Sketris.We also thank all of the physicians and patients who participated in the survey.
Contributors: BM and MLB contributed to all aspects of study planning, design, analysis, and reporting. RLK, AK, and KB contributed to interpretation of the data, review of drafts the manuscript and, to a lesser extent, study planning, questionnaire design, and data collection. JL and RGE contributed to study design, interpretation, and review of the manuscript. RP contributed to US components of study design and organised and supervised data collection and entry in Sacramento. SAM contributed to the analysis plan and interpretation and to subsequent discusssion of these components of the manuscript. BM and MLB are guarantors.
Funding Health Transition Fund, Health Canada. Barbara Mintzes also received a PhD training fellowship for this research from the National Health Research and Development Programme (NHRDP) and the Canadian Institutes of Health Research.
Competing interests None declared.
Like many journals the BMJ derives income from pharmaceutical advertising. Direct to consumer advertising of prescription drugs competes with this channel.