It is believed that pharmaceutical industry sponsorship of clinical research leads to the development of multiple ties between clinicians and the pharmaceutical industry. To quantify this relationship we conducted a survey of medical specialists listed in the Medical Directory of Australia in 2002 and 2003.
A questionnaire was mailed that elicited information about all aspects of research relationships between clinicians and pharmaceutical companies. The odds of reporting multiple additional ties (financial and professional) with pharmaceutical companies by clinicians who had an active research relationship were compared with those who did not. All clinicians who returned a completed questionnaire about their research activities were included in the study.
A questionnaire was mailed to 2120 medical specialists; 823 (39%) responded. Of these, 338 (41%) reported involvement in industry-sponsored research in the previous year. They were more likely than others to have been offered industry-sponsored items or activities valued at more than AU $500 (>US $382; odds ratio [OR], 3.5; 95% confidence interval [CI], 2.6-4.7) and support for attending international conferences (OR, 5.4; 95% CI, 3.9-7.4). The strongest associations were seen for acting as a paid consultant to industry (OR, 9.0; 95% CI, 3.9-20.4) and for membership on advisory boards (OR, 6.9; 95% CI, 5.1-9.6). There was a strong relationship between research collaboration and accumulation of industry ties. For 1 additional tie the OR was 2.2 (95% CI, 1.2-3.8) and rose to 6.3 (95% CI, 3.5-11.1) with 3 ties and 41.8 (95% CI, 14.5-143.4) with 6 or more ties.
Medical specialists who have research relationships with the pharmaceutical industry are much more likely to have multiple additional ties than those who do not have research relationships. Institutional review should discourage clinical researchers from developing multiple ties.
Interactions between medical specialists and pharmaceutical manufacturers are important for the development and evaluation of new medicines. However, interaction with industry has the capacity to create conflicts of interest. While these risks are well recognized, studies have shown that physicians underestimate the potential for different interactions to cause problems. For instance, many do not consider that acceptance of small gifts creates an ethical dilemma.1-3 Interaction with industry in one area may lead to the development of other ties. The development of multiple ties may blur professional boundaries and responsibilities. Research collaboration, an important and growing area of engagement between industry and clinical researchers, may lead to other significant relationships, such as advisory panel membership, payment for consultation to the industry, and substantial recompense to attend international conferences.4-6 These ties may create a sense of collegiality, and the resulting obligation and need to reciprocate may not be consciously felt.2,7-9 The term “entanglement” has been used to describe this process, but there has been little direct research into the development of multiple ties.10,11
As part of a study of the relationship between Australian medical specialists and the pharmaceutical industry,12 we explored the extent to which participation in pharmaceutical industry–sponsored research by medical specialists is associated with the accumulation of multiple additional ties with the pharmaceutical industry. We included private practitioners in this research because of evidence that they have become increasingly involved in commercially sponsored research.4
We surveyed medical specialists listed in the 2002 and 2003 editions of the Medical Directory of Australia (Australasian Medical Publishing Co Proprietary Ltd, Strawberry Hills, New South Wales, Australia). Five thousand names supplied by the publishers of the Medical Directory of Australia were stratified according to clinical subspecialty and the state in which the practice address was located. Specialists with minimal prescribing responsibility (eg, surgeons and anesthetists) were excluded. The questionnaire sought information on all aspects of specialists’ relationship with the pharmaceutical industry. Respondents were asked about general interactions with industry, such as attendance at industry-sponsored medical educational meetings, sponsored symposiums, and drug product launches. General interactions were not classified as ties because they do not directly involve reciprocity. Respondents were also asked about more specific relationships with industry through medical or scientific advisory panel membership, acting as an expert speaker about a pharmaceutical product, assistance in the preparation of industry-sponsored professional educational materials, payment for consultation, and shareholding. We also asked for details of offers they had received of honoraria, gifts, and travel or accommodation support.
Respondents were categorized as having an active research relationship with the pharmaceutical industry if they reported any participation in industry-sponsored research in the previous 12 months. Respondents with and without an active research relationship were compared for the frequency with which they reported 1 or more additional ties with the pharmaceutical industry during the same period. Ties were defined as specific interactions that were likely to involve a degree of reciprocity. Data were analyzed using SAS software (SAS Institute Inc, Cary, NC). Associations between research relationships and industry ties were analyzed by means of the odds ratio (OR) and 95% confidence interval (CI). We stratified the numbers of additional ties and analyzed the associations by performing a test for linear trend.
The questionnaire was mailed to 2253 specialists. In 133 instances, we received a reply that the individual was ineligible (deceased, emigrated, or retired). Of the remaining 2120 specialists, 823 (39%) completed and returned the survey instrument. The Medical Directory of Australia relies on members updating their own data; thus, it is likely that several nonrespondents had also left active practice (information provided by the Medical Directory of Australia).
Characteristics of respondents
The respondents were similar to the original sample for geographic location and clinical specialty.12 The mean (SD) age of the respondents was 48 (10) years; most respondents (79%) were men. Respondents spent 46% of their time in salaried hospital practice and 47% in private clinical practice, and, on average, saw 20 patients and wrote 20 prescriptions daily (Table 1).
General interactions with the pharmaceutical industry
Reports of general interactions with industry in the previous year were common: 709 respondents (86%) reported seeing a sales representative; 518 (63%) had attended an industry-sponsored medical educational meeting; 419 (51%) had attended a product-specific sponsored symposium, and 342 (41%) had attended a drug product launch. Offers of gifts of little monetary value for personal use, such as chocolates or wine, were reported by 420 respondents (51%) (Table 1).
Specific interactions with the pharmaceutical industry (ties)
In the previous year, 193 respondents (23%) had been a member of a medical or scientific advisory panel; 131 (16%) had acted as an expert speaker about a pharmaceutical product; 133 (16%) had been involved in the preparation of industry-sponsored professional educational materials; 46 (5%) had acted as a paid consultant for industry, and 56 (7%) held shares in a pharmaceutical company (Table 1). Of the 144 respondents who received payment for advisory panel membership or consulting, 121 (84%) estimated this to be less than AU $5000 (US $3820)1 annually; an income of between AU $5000 and $10 000 (US $3820-$7640) was declared by 16 respondents, and 7 reported an income of more than AU $10 000 (US $7640). Offers of industry support for items and activities valued at more than AU $500 (US $382) were reported by 418 respondents (51%); support for attendance at a national conference was reported by 328 respondents (40%); and 259 respondents (31%) reported offers of support for attending international conferences (Table 1).
Associations between industry research sponsorship and accumulation of additional ties with the pharmaceutical industry
Three hundred thirty-eight specialists (41%) reported participating in industry-sponsored medical research in the previous year. They were slightly younger, more likely to be men, and more likely to spend more than 20% of their time in research than those with no research relationship with industry (Table 2). Respondents with research relationships were more likely than those without such relationships to report offers of support for attendance at national conferences (OR, 1.7; 95% CI, 1.3-2.3); offers of industry-sponsored items or activities valued at more than AU $500 (US $382) (OR, 3.5; 95% CI, 2.6-4.7); offers of support for attending international conferences (OR, 5.4; 95% CI, 3.9-7.4); involvement in the preparation of industry-sponsored professional educational materials (OR, 6.0; 95% CI, 3.6-9.2); and acting as a paid consultant for industry (OR, 9.0; 95% CI, 3.9-20.4). However, there was no significant difference between active researchers and others for holding shares in a pharmaceutical company (OR, 1.3; 95% CI, 0.8-2.3).
A strong gradient was observed in the associations between research collaboration with industry and accumulation of additional ties. For 1 additional tie the OR (those with industry research relationships vs those without) was 2.2 (95% CI, 1.2-3.8) and rose to 6.3 (95% CI, 3.5-11.1) with 3 ties and to 41.8 (95% CI, 14.5-143.4) with 6 or more ties (Table 3). A test for linear trend yielded a positive result (P<.001).
Research collaboration between medical specialists and the pharmaceutical industry is widespread and strongly associated with multiple additional ties. In the univariate analyses, associations were strongest for offers of high value (eg, overseas travel) and for significant professional engagement (eg, advisory panel membership and payment for consultation). The most impressive finding was the steepness of the gradient in the relationship between research collaboration and increasing numbers of additional ties of any type.
The increased ORs for contributions of higher value and the steep gradient of ORs for multiple ties may indicate that clinical researchers are particularly valued by the pharmaceutical industry and that this leads to the development of multiple ties. However, our data do not enable us to draw confident conclusions about the direction of these relationships or their cause. It is also possible that industry engagement leads some clinicians to become involved in research.
Regardless of chronology or causal link, the association between research collaboration and holding multiple other ties with industry raises questions about the capacity of physicians to remain independent and obligation free. These questions are particularly pertinent to medical specialists because they are likely to be opinion leaders who can influence not only their peer groups but also general practitioners and undergraduate and postgraduate students. Other studies1-3 have observed that specialists participating in industry-sponsored research express confidence in their capacity to remain impervious to industry influence. This confidence is at odds with the considerable literature on the unconscious and unintentional processes through which obligation may be established.2 Given what is known about reciprocity and the “gift relationship,” each additional tie with industry potentially compounds the relationship and increases the potential for obligation, entanglement, and conflicts of interest.
Substantial financial gain does not seem to be the compelling motivation in these relationships, at least among Australian physicians. Shareholding was uncommon and was not associated with involvement in commercially funded research. The level of fees received for consultation or advisory panel membership was also generally modest (typically <AU $5000 [<US $3820] per year). However, financial interests are not the only, or necessarily the most powerful, secondary interests that determine physicians’ attitudes and behavior. Indeed, an examination of the history of the profession suggests that other, nonpecuniary interests, including social status, professional power, academic advancement, desire for professional recognition, and need to achieve success in research-funding applications, may all be more important to physicians than gifts or entertainment.13-16 This may explain both the effect that participation on advisory boards and as consultants has on physicians’ attitudes toward engagement with the pharmaceutical industry and the increasing recruitment of physicians to these functions by the pharmaceutical industry.15
The findings of our survey are limited by the low response rate (<40%, although this rate is consistent with other surveys on similar issues16), reliance on self-report, and the focus on a specific population. While our respondents were similar for age, sex, state of practice, and clinical specialty to those in the original sampling frame, nonresponders may have differed from responders in their views.12 Reliance on self-report leaves open the possibility of selective or inaccurate recall, although anonymity may have provided protection. The focus on medical specialists limits generalization to other medical populations, such as primary care physicians. However, we believe it unlikely that selection bias could account for the strong associations and the steep gradient seen in our results, and the large sample size gives us confidence in the precision of our estimates.
The changing face of commercially sponsored research has increased the opportunity for individual physicians practicing in the community to engage in industry-sponsored research and to advance their own professional and clinical interests. However, the benefits from legitimate interactions with industry must be weighed against the possibility of creating extensive dualities of interest and possible conflict of interests. At the institutional level, review boards and professional organizations must be aware that clinical research relationships between physicians and the pharmaceutical industry are predictive of additional ties and require careful attention to minimize potential conflicts of interest. At the individual level, physicians should be mindful of ties that may proliferate and bind.
Correspondence: David Henry, MB, ChB, Level 5 Clinical Sciences Building, Newcastle Mater Hospital, Waratah, New South Wales 2298, Australia (firstname.lastname@example.org).
Accepted for Publication: July 3, 2005.
Financial Disclosure: Dr Henry has in the last 3 years held a contract with Wyeth to review the toxicity of nonsteroidal anti-inflammatory drugs. As a member of the Australian Pharmaceutical Benefits Advisory Committee, he was a respondent in legal actions taken by pharmaceutical companies against the committee in the Australian Federal Court. Dr Day serves as an Advisory Board member for MSD Australia (rofecoxib, etoricoxib), MSD Asia (rofecoxib), Abbott Australia (adalimumab), Schering-Plough Australia (infliximab), Amgen Australia (anakinra), GlaxoSmithKline Consumer Australia (paracetamol), and, previously, Pfizer Australia (celecoxib). Any honoraria for these activities are placed in audited trust funds of St Vincent’s Hospital, Sydney, Australia, to be used to support academic activities within the Discipline of Clinical Pharmacology.
Funding/Support: This study was supported by a grant from the National Health and Medical Research Council of Australia, Canberra.
Acknowledgment: We are grateful to the medical specialists who responded to our survey.
JF Towards an understanding of the ethics of pharmaceutical industry gift-giving. Am J Bioeth
2003;339- 46PubMedGoogle ScholarCrossref
SR Do doctors rely on pharmaceutical industry funding to attend conferences and do they perceive that this creates bias in their drug selection? results from a questionnaire survey. Pharmacoepidemiol Drug Saf
2003;12663- 667PubMedGoogle ScholarCrossref
SS Conflict of interests, conflicting interests and interesting conflicts: part 3. J Clin Ethics
1996;7184- 186PubMedGoogle Scholar
R Who pays for the pizza? redefining the relationships between doctors and drug companies, 1: entanglement. BMJ
2003;3261189- 1192PubMedGoogle ScholarCrossref
LA Implementation of financial disclosure policies to manage conflict of interest. Health Aff (Millwood)
2004;23206- 214PubMedGoogle ScholarCrossref
M The Birth of the Clinic. Sheridan Smith
AMtrans New York, NY Vintage Books1973;
C Pharma goes to the laundry: public relations and the business of medical education. Hastings Cent Rep
2004;3418- 23PubMedGoogle ScholarCrossref
LA Conflicts of interest in academic research: policies, process, and attitudes. Account Res
2004;1183- 102PubMedGoogle ScholarCrossref