Physician Preferences and Attitudes Regarding Industry Support of CME Programs

Abstract

Pharmaceutical and other health care-related companies spend approximately $12 to $15 billion per year ($8000-$15000 per year, per physician) on marketing.1Blumenthal D. Doctors and drug companies.N Engl J Med. 2004; 351: 1885-1890Crossref PubMed Scopus (194) Google Scholar One marketing approach used by many pharmaceutical companies is to provide financial support of continuing medical education (CME) programs.1Blumenthal D. Doctors and drug companies.N Engl J Med. 2004; 351: 1885-1890Crossref PubMed Scopus (194) Google Scholar, 2Relman A.S. Separating continuing medical education from pharmaceutical marketing.JAMA. 2001; 285: 2009-2012Crossref PubMed Scopus (149) Google Scholar, 3Wazana A. Physicians and the pharmaceutical industry: Is a gift ever just a gift?.JAMA. 2000; 283: 373-380Crossref PubMed Scopus (1193) Google Scholar In recent years, this support has increased. Ten years ago, 17% of CME funding came from industry; today, that number is 40%.4Packer S. Parke II, D.W. Ethical concerns in industry support of continuing medical education: the con side.Arch Ophthalmol. 2004; 122: 773-776Crossref PubMed Scopus (6) Google Scholar, 5Leichter S.B. Continuing medical education in diabetes: the impending crisis (the business of diabetes).Clin Diabetes. 2004; 22: 174-177Crossref Google Scholar Between 1992 and 2001, industry support of medical school-sponsored CME quintupled.6Davis D.A. CME and the pharmaceutical industry: Two worlds, three views, four steps.Can Med Assoc J. 2004; 171: 149-150Crossref Scopus (13) Google Scholar Organizations that conduct CME programs claim that without financial support from industry, programs must rely on registration fees, which, when combined with travel expenses, would make the programs unaffordable for many participants.2Relman A.S. Separating continuing medical education from pharmaceutical marketing.JAMA. 2001; 285: 2009-2012Crossref PubMed Scopus (149) Google Scholar Physicians attend CME programs for many reasons, including fulfilling state medical licensure requirements, maintaining hospital privileges and specialty society memberships, and obtaining new knowledge and skills.7Wilson F.S. Continuing medical education: ethical collaboration between sponsor and industry.Clin Orthop Relat Res. 2003; 412: 33-37Crossref PubMed Scopus (12) Google Scholar, 8Holmer A.F. Industry strongly supports continuing medical education.JAMA. 2001; 285 (erratum in: JAMA. 2001;285:2451): 2012-2014Crossref PubMed Scopus (49) Google Scholar Many physicians also regard CME courses as their most valuable source for clinical information.7Wilson F.S. Continuing medical education: ethical collaboration between sponsor and industry.Clin Orthop Relat Res. 2003; 412: 33-37Crossref PubMed Scopus (12) Google Scholar However, evidence suggests that CME programs sponsored by industry not only may be more biased (in favor of the sponsoring companies’ products) than programs not sponsored by industry3Wazana A. Physicians and the pharmaceutical industry: Is a gift ever just a gift?.JAMA. 2000; 283: 373-380Crossref PubMed Scopus (1193) Google Scholar, 5Leichter S.B. Continuing medical education in diabetes: the impending crisis (the business of diabetes).Clin Diabetes. 2004; 22: 174-177Crossref Google Scholar but also may influence physicians’ professional behavior (eg, increased prescriptions of the sponsor’s medication).3Wazana A. Physicians and the pharmaceutical industry: Is a gift ever just a gift?.JAMA. 2000; 283: 373-380Crossref PubMed Scopus (1193) Google Scholar, 4Packer S. Parke II, D.W. Ethical concerns in industry support of continuing medical education: the con side.Arch Ophthalmol. 2004; 122: 773-776Crossref PubMed Scopus (6) Google Scholar, 5Leichter S.B. Continuing medical education in diabetes: the impending crisis (the business of diabetes).Clin Diabetes. 2004; 22: 174-177Crossref Google Scholar These findings raise the ethical concern of industry influence on physicians who participate in CME programs. Each year, Mayo Clinic College of Medicine School of CME (Rochester, Minn) conducts comprehensive internal medicine CME courses, some of which receive financial support from industry. Over the years, directors of these courses have received mixed feedback from participants regarding industry support. To better understand physician preferences and attitudes regarding industry support of CME programs, an anonymous survey of physicians attending 4 Mayo Clinic internal medicine CME courses was conducted in 2004. Pharmaceutical and other health care-related companies financially support 2 of the courses (Mayo Internal Medicine Board Review and Mayo Clinical Reviews) with unrestricted grants and in accordance with Accreditation Council for Continuing Medical Education (ACCME) guidelines. Financial support is given directly to the course organizers and is used to defray the costs of the courses; the companies have no role in planning or conducting the courses. These courses allow industry-sponsored exhibits adjacent to the meeting room. The other 2 courses (Mayo Selected Topics in Internal Medicine and Mayo Practice of Internal Medicine) are not supported by industry and have no exhibitors. Respondents completed the survey during a break period in the middle of each course. The one-page survey instrument asked participants their age, sex, and years in practice. The survey also included 4 specific questions regarding industry support of CME activities. The results of the survey comprise the data set of this study. Between-group responses to questions were compared using the Pearson chi-squared test. Ordinal logistic regression was used to fit multivariate models. The response ordering was “yes,” “no,” and “no preference” (or equivalent). A P value less than .05 was considered significant. All analyses were conducted using JMP 4.0.4 software (SAS Institute, Inc., Cary, NC). Permission to perform an analysis of the surveys was granted by the Mayo Clinic Institutional Review Board in accordance with federal regulations. Of 1603 physicians attending the courses, 1130 (70.5%) completed the survey. Most of the 1603 attended only one course; only 19 (1.2%) attended 2 courses, and none attended 3 or 4 courses. (It is unknown whether the physicians who attended 2 courses completed the survey once, twice, or at all). Of the 1130 survey respondents, 671 (59.4%) attended a course sponsored by industry. Table 1 highlights the characteristics of the participants who completed the survey. In the multivariate analyses, age and years in practice were found to highly relate to each other. Therefore, the rest of the article only presents data related to years in practice.Table 1Characteristics of 1130 Physician Attendees Who Completed the SurveyCharacteristicNo. of Respondents (%)Sex1015 (89.8) Male786 (77.4) Female229 (22.6)Age, years1121 (99.2) ≤40189 (16.9) 41-50331 (29.5) 51-60324 (28.9) ≥61277 (24.7)Years in practice1030 (91.2) 0-10204 (19.8) 11-20288 (28.0) 21-30289 (28.1) ≥31249 (24.1)Type of course attended1130 (100) Not supported by industry459 (40.6) Supported by industry671 (59.4) Open table in a new tab In response to the question “What type of CME course (industry supported or not) do you prefer to attend?”, 58.3% of respondents indicated no preference (Table 2). Among those physicians who indicated a course preference, the majority preferred non-industry-supported courses. Responses to this question differed significantly by years in practice (P <.001). Experienced physicians (in practice more than 30 years) preferred industry-supported courses more than less experienced physicians. Responses also differed significantly by course attended (P <.001). More participants of non-industry-supported courses preferred that type of course than participants of industry-supported courses. In the multivariate model, more than 30 years in practice (as opposed to 10 or fewer years in practice) and attending an industry-supported course were independent predictors of response to this question (P <.001 for both).Table 2Responses to Question 1—Preferred Course to AttendCharacteristic of RespondentsCourse Preference by PercentageP Value⁎Pearson chi-squared test.Industry-supportedNon-industry-supportedNo PreferenceAll respondents (n = 1121)8.433.358.3Sex (n = 1009).14 Male (n = 783)8.833.058.2 Female (n = 226)4.935.859.3Years in practice (n = 1022)<.001 0-10 (n = 202)4.040.155.9 11-20 (n = 285)6.336.157.5 21-30 (n = 287)7.038.055.1 ≥31 (n = 248)†Significant predictor of response in the multivariate model.16.515.767.7Type of course attended (n = 1121)<.001 Not supported by industry (n = 456)5.945.648.5 Supported by industry (n = 665)†Significant predictor of response in the multivariate model.10.124.865.1 Pearson chi-squared test.† Significant predictor of response in the multivariate model. Open table in a new tab In regard to preference for a CME course with or without exhibitors (question 2), one-half of the respondents indicated no preference, whereas the other one-half was evenly divided (24.8% each) in their preference for courses with or without exhibitors (Table 3). Responses to this question differed significantly by sex and years in practice. More men preferred courses with exhibitors, whereas more women preferred courses without exhibitors (P = .005). Compared with more experienced respondents, physicians in practice fewer than 31 years preferred courses without exhibitors (P <.001). Responses also differed significantly by course attended (P <.001). More participants of non-industry-supported courses preferred courses without exhibitors. In the multivariate model, more than 20 years in practice (as opposed to 10 or fewer years in practice) and attending an industry-supported course were independent predictors of response to this question (all P <.001). Sex was not a predictor in the multivariate model.Table 3Responses to Question 2—Preference for ExhibitorsCharacteristic of RespondentsCourse Preference by PercentageP Value⁎Pearson chi-squared test.With ExhibitorsWithout ExhibitorsNo PreferenceAll respondents (n = 1121)24.824.850.4Sex (n = 1108).005 Male (n = 781)26.622.451.0 Female (n = 227)18.931.749.3Years in practice (n = 1022)<.001 0-10 (n = 202)20.329.250.5 11-20 (n = 286)17.830.851.4 21-30 (n = 285)†Significant predictor of response in the multivariate model.17.225.357.5 ≥31 (n = 249)†Significant predictor of response in the multivariate model.45.811.243.0Type of course attended (n = 1121)<.001 Not supported by industry (n = 456)15.143.941.0 Supported by industry (n = 665)†Significant predictor of response in the multivariate model.31.411.756.8 Pearson chi-squared test.† Significant predictor of response in the multivariate model. Open table in a new tab In response to question 3, “Do you believe CME courses should accept industry support if doing so reduces the overall cost of the course?”, 62.3% of respondents answered “yes” (Table 4). Responses differed significantly by sex, with more women answering “no” or “no preference” (P = .02), and by years in practice, with increasing years of experience directly associated with answering “yes” (P = .001). Responses to this question also differed significantly by course attended (P <.001). A majority of industry-supported course attendees (71.5%) answered “yes,” whereas the attendees of the non-industry-supported courses were more evenly divided (48.7% answered “yes,” 33.1% answered “no”). However, in the multivariate model, only attending an industry-supported course was an independent predictor of response to this question (P <.001).Table 4Responses to Question 3—Should Industry Support Be Used to Reduce Costs?Characteristic of RespondentsResponse by PercentageP Value⁎Pearson chi-squared test.YesNoNo PreferenceAll respondents (n = 1126)62.323.414.3Sex (n = 1011).02 Male (n = 784)64.521.913.5 Female (n = 227)54.227.818.1Years in practice (n = 1027).001 0-10 (n = 204)55.428.915.7 11-20 (n = 286)60.127.312.6 21-30 (n = 288)62.922.614.6 ≥31 (n = 249)72.313.314.5Type of course attended (n = 1126)<.001 Not supported by industry (n = 456)48.733.118.2 Supported by industry (n = 670)†Significant predictor of response in the multivariate model.71.516.911.6 Pearson chi-squared test.† Significant predictor of response in the multivariate model. Open table in a new tab For the final question, “Is it your impression that the contents of CME courses supported by industry tend to be biased in favor of the supporting companies?”, 53% of participants responded “no” (Table 5). Responses to this question differed significantly by sex, years in practice, and course attended. More men than women answered “no” (P = .01). Compared with physicians in practice longer than 30 years, a greater percentage of less experienced respondents answered “yes” (P <.001). A majority of non-industry-supported course participants (52.7%) answered “yes,” whereas a majority (63.7%) of the participants of industry-supported courses answered “no” (P <.001). In the multivariate model, only 21 to 30 years in practice and attending an industry-supported course were independent predictors of response to this question (P <.001 each).Table 5Responses to Question 4—Are Industry-Supported CME Courses Biased?Characteristic of RespondentsResponse by PercentageP Value⁎Pearson chi-squared test.YesNoNo OpinionAll respondents (n = 1120)35.953.011.1Sex (n = 1006).01 Male (n = 782)33.155.811.1 Female (n = 224)41.544.613.8Years in practice (n = 1024)<.001 0-10 (n = 202)43.641.614.9 11-20 (n = 287)33.556.510.1 21-30 (n = 287)†Significant predictor of response in the multivariate model.41.850.28.0 ≥31 (n = 248)24.261.714.1Type of course attended (n = 1120).001 Not supported by industry (n = 457)52.737.69.6 Supported by industry (n = 663)†Significant predictor of response in the multivariate model.24.363.712.1 Pearson chi-squared test.† Significant predictor of response in the multivariate model. Open table in a new tab This survey was designed to assess physician preferences and attitudes regarding industry support of CME programs. A majority of participants indicated no preference for industry-supported or non-industry-supported courses, and one-half indicated no preference for courses with or without exhibitors. A majority believed CME courses should accept industry support to reduce the cost of courses, and about one-half thought that courses supported by industry were not biased. Organizers of CME programs should find these results interesting and helpful in planning future courses. Results indicate that the preferences and attitudes regarding industry support of CME programs vary significantly according to the number of years in practice and type of course attended (industry supported or not). Compared with less experienced physicians, those in practice more than 30 years were more likely to prefer industry-supported courses and courses with exhibitors and believed CME courses should accept industry support. Several explanations most likely account for these findings. Physicians who have been in practice for many years have had more opportunities to attend industry-supported CME programs and interact with industry representatives and may believe they are not influenced by these interactions. Also, these physicians may enjoy the interactions with industry representives and derive educational value for themselves and benefits for their patients (such as drug samples) from industry-supported CME programs. In contrast, physicians in practice fewer years, who were more likely to prefer non-industry-supported courses, may be more skeptical of industry-sponsored CME programs. These attitudes may partly be a result of recent efforts by medical schools, residency programs, professional societies, and organizations,9Coyle S.L. Ethics and Human Rights CommitteeAmerican College of Physicians-American Society of Internal MedicinePhysician-industry relations Part 2: organizational issues.Ann Intern Med. 2002; 136: 403-406Crossref PubMed Scopus (92) Google Scholar, 10American Medical Association, Council on Ethical and Judicial AffairsCode of Medical Ethics: Current Opinions with Annotations. AMA Press, Chicago, IL2004Google Scholar such as No Free Lunch,11No Free Lunch. Available at: http://www.nofreelunch.org. Accessed January 30, 2006.Google Scholar to raise awareness of potential conflicts of interest in physician–industry interactions. Responses also varied significantly according to type of course attended. More physician respondents who attended non-industry-supported courses than industry-supported courses believed that industry bias exists. Several reasons may account for this finding. Some participants of non-industry-supported CME programs may consciously avoid industry-supported programs, and some respondents likely took into account the type of course they were attending and answered the questions to justify their choice. A majority of the survey respondents—including nearly one-half the participants of non-industry-supported courses—believed CME courses should accept industry support if doing so reduces the overall cost of the course. Although it has been reported that gifts, regardless of the size, may instill in the physician recipient a sense of obligation to reciprocate,1Blumenthal D. Doctors and drug companies.N Engl J Med. 2004; 351: 1885-1890Crossref PubMed Scopus (194) Google Scholar, 12Dana J. Loewenstein G. A social science perspective on gifts to physicians from industry.JAMA. 2003; 290: 252-255Crossref PubMed Scopus (441) Google Scholar it is unknown whether industry-defrayed costs of attending CME programs engender a similar obligation. Regardless of physician preferences and attitudes, financial support of CME programs by industry is likely to continue. How should concerns about industry support of CME be addressed? First, CME organizers should explicitly acknowledge that the primary function of CME is to improve the welfare of patients by enhancing the knowledge and skills of physicians,6Davis D.A. CME and the pharmaceutical industry: Two worlds, three views, four steps.Can Med Assoc J. 2004; 171: 149-150Crossref Scopus (13) Google Scholar not to facilitate physician–industry interactions. Several professional organizations, including the American Medical Association,10American Medical Association, Council on Ethical and Judicial AffairsCode of Medical Ethics: Current Opinions with Annotations. AMA Press, Chicago, IL2004Google Scholar the ACCME,13Accreditation Council for Continuing Medical Education. Available at: http://www.accme.org. Accessed January 30, 2006.Google Scholar and the American College of Physicians,9Coyle S.L. Ethics and Human Rights CommitteeAmerican College of Physicians-American Society of Internal MedicinePhysician-industry relations Part 2: organizational issues.Ann Intern Med. 2002; 136: 403-406Crossref PubMed Scopus (92) Google Scholar have made recommendations regarding industry support of CME programs. Taken together, these recommendations can be summarized as follows:•Industry support must be completely unrestricted;•All CME faculty conflicts of interest must be declared before the program begins;•The industry sponsor should have no role in the planning or evaluation of program content;•The topics should be presented without bias, particularly if the products of the industry sponsor are discussed; and•Support should not be given to participants but to the program organizers to reduce registration fees. Adherence to these recommendations should prevent most inappropriate industry influence or bias in CME programs. This study has several limitations. Although the number of respondents was large and the response rate excellent, the survey itself was brief, which precluded an in-depth examination of physician attitudes regarding industry support of CME programs. In addition, participants of courses organized and conducted by Mayo Clinic College of Medicine School of CME, Rochester, Minn, were the only group surveyed. However, physicians attend these courses from nearly every state in the United States and a number of different countries. Nevertheless, results may not be generalizable to all physicians (such as individuals who do not attend courses organized and conducted by Mayo Clinic or physicians who attend noninternal medicine courses). Also, the survey assessed participant preferences and attitudes regarding industry support of CME programs, not the actual effects of industry support on programs. The authors thank Kelley M. Sandvik for her assistance in gathering data. Editing, proofreading, and reference verification were provided by the Section of Scientific Publications, Mayo Clinic.