Chimonas, Susan PhD; Evarts, Susanna D.; Littlehale, Sarah K.; Rothman, David J. PhD
Over the last 15 years, clinical conflicts of interest (CCOIs) have received unprecedented attention in the United States. Since the 1990s, a substantial body of academic research has demonstrated the power of industry gifts and payments to influence physicians’ beliefs and practices.1–6 State and federal regulators have successfully prosecuted high-profile cases in which companies induced physicians to prescribe or promote their products.7–9 Media coverage has been extensive, exposing the potential for company largesse to influence patient care.10–13
Several prominent medical and academic organizations have promulgated recommendations for managing CCOIs. A clear consensus has emerged. The Institute of Medicine (IOM) in 2009, the Association of American Medical Colleges (AAMC) in 2008, and the ABIM Foundation (ABIMF) jointly with the Institute on Medicine as a Profession (IMAP) in 200614–16 all recommend eliminating industry gifts and meals, speakers’ bureaus, and ghostwriting; they also urge strict control of industry payments for consulting, educational grants, and honoraria. The differences among the recommendations are minor. The ABIMF–IMAP, for example, would “prohibit” speakers’ bureaus; the IOM and the AAMC would “strongly discourage” them. The recommendations have won support in the medical profession and in the media, including editorial endorsement by the New York Times.17
IMAP has been analyzing the degree to which medical schools’ policies are consistent with these recommendations. In 2008, IMAP first systematically collected and evaluated medical schools’ CCOI policies.18 At that time, few schools had implemented the recommended policies; a majority lacked sufficiently stringent policies in most areas. More than a quarter had enacted no policies in any area. Many, however, indicated that they were in the process of revising their policies.18
To examine how much progress has been made, in 2011, IMAP again collected and analyzed medical schools’ CCOI policies. The findings reveal an extraordinary transformation over a three-year period, with many more stringent policies in place now than in 2008. Without question, however, room for improvement remains.
We hypothesized that public medical schools were more likely than private ones to have stringent CCOI policies because of the existence of state laws governing conflicts of interest at public institutions. Similarly, we hypothesized that schools that owned their hospitals would have more stringent policies than would hospital-affiliated schools, as CCOIs can lead to more costly care. We also hypothesized that more research-intensive institutions, as assessed by National Institutes of Health (NIH) funding, would be more likely to have stringent policies because of the attention given to conflicts of interest by state and federal bodies. We also wanted to determine whether any of these independent variables were correlated with changes in policy strength from 2008 to 2011.
We sought to collect medical schools’ policies in the CCOI areas identified as salient by the AAMC, IOM, and ABIMF–IMAP. IMAP’s 2008 study found relatively few policies online, so that study did not pursue this method of data collection. By 2011, however, most schools were posting policies on the Web. Accordingly, beginning in July 2011, we searched online for the policies of all 133 MD-granting U.S. medical schools in existence at that time, using these key words:
* Policies, faculty handbook, faculty policies, resident policies
* Conflict of interest, clinical conflict of interest, conflict of interest policy
* Industry interactions, industry relationships, industry relations policy
* Vendor interactions, vendor relations, vendor policy, vendor access, vendor guide
* Ghostwriting, authorship, authorship guidelines
* Samples, pharmaceutical samples
* Pharmaceutical and medical device manufacturers
* Outside professional activities, consulting, outside employment
* Compliance office, compliance
* Pharmacy and therapeutics, formulary, procurement, recusal
* Speakers’ bureaus, industry sponsored speaking
* Continuing medical education, CME, ACCME
* Gifts, gifts policy
* Industry-sponsored meals
* Ethics policy, ethical guidelines
When no policies were found online (as was the case for seven schools), we also contacted those schools directly to obtain their policies. Individuals responsible for CCOI management (e.g., deans, compliance officers) received e-mails describing the project and requesting all relevant policies. Those who did not respond received follow-up e-mails. A final round of follow-ups occurred by telephone. One of the seven schools provided policies before data collection ended on September 30, 2011.
The study received approval from the institutional review board of Columbia University. No financial incentives were offered for participation.
Policy coding and analysis
We used a coding system to gauge the strength of schools’ CCOI policies, based on the recommendations issued by the ABIMF-IMAP, AAMC, and IOM (see Supplemental Digital Table 1, which can be accessed at http://links.lww.com/ACADMED/A153). These proposed that institutions (1) eliminate industry gifts, meals, and ghostwriting; (2) prohibit or “strongly discourage” speakers’ bureaus; (3) establish central repositories for product samples and industry funds for continuing medical education (CME), scholarships, fellowships, and travel; (4) require that members of pharmacy and therapeutics (P&T) and other purchasing committees be free of conflicts of interest; and (5) require full transparency for industry honoraria and consulting contracts.
Our codebook, an earlier version of which was used in the 2008 IMAP study, provides specific criteria for assessing the strength of policies. These criteria were identified by two of us (S.C. and D.J.R.) in 2006–2007 through intensive fieldwork at more than a dozen medical schools.19–21 We updated the 2008 codebook to capture important changes in policy language and scope and to ensure consistent coding. For example, travel and scholarships/fellowships were split into two categories (they were a single category in 2008) because many more schools’ policies now address them separately. To ensure accurate comparisons between the 2008 and 2011 data, we recoded all 2008 policies using the updated codebook.
For all but two policy areas, the codebook specifies four codes: 0 (no policy), 1 (permissive), 2 (moderate), and 3 (stringent). CME and ghostwriting are coded on a scale of 0, 1, and 3, as no appreciable difference between moderate and permissive ratings exists in these categories. (See the Supplemental Digital Appendix 1 for the complete codebook. It can be accessed at http://links.lww.com/ACADMED/A153.)
To assess intercoder reliability, a random sample of 17% of the policies was coded independently by two of us (S.D.E. and S.K.L.). Differences in coding scores were resolved through discussion. The remaining policies were divided in half and coded by either S.D.E. or S.K.L.
We also assessed policy strength at the institutional level, examining schools’ scores across the 12 CCOI areas. Once all policies were coded, we averaged each school’s 12 policy scores (weighted equally) to create a measure of overall strength.
We then compared the data from 2008 and 2011, assessing changes in the frequency of stringent, moderate, permissive, and “no policy” scores in each CCOI area. We also evaluated changes in schools’ overall policy strength.
We coded the following institutional characteristics:
* Institution type: coded as 0 (private) or 1 (public)
* Hospital relationship: coded as 3 (hospital/hospitals owned by medical school), 2 (hospital/hospitals affiliated with but not owned by medical school), or 1 (mixture of ownership and affiliation across multiple hospitals)
* NIH funding status: dollar amount of NIH funding in FY 2007 and 2010
The NIH figures include only grant and contract funding. However, to our knowledge, no more complete index of medical schools’ public research funding exists.
We used one-way ANOVA and regression tests to assess whether these variables (public versus private, hospital ownership versus affiliation, and amount of NIH funding) were associated with differences in CCOI policy strength. A P value of .05 was used to test significance. We used Excel Data Analysis ToolPak statistical software, 2010 version (Microsoft Corp., Redmond, Washington), along with XLSTAT-Pro, 2012.6.01 version (Addinsoft, New York, New York), to conduct our analysis.
Our data collection yielded policies for 127 (95%) of the 133 accredited MD-granting medical schools that existed as of September 30, 2011. We found no statistically significant differences between the institutional characteristics of those 127 schools and the 6 schools that did not provide data. Eighty-four schools (66%) from our sample were public institutions compared with 85 (64%) of the overall population. Seventy-one schools (56%) owned one or more of their associated hospitals, comparable to 76 (57%) of all medical schools. In FY 2010, the schools in our sample received an average of $94,307,363 in NIH funding (median $50,365,766) compared with $91,190,931 (median $49,869,302) among all medical schools.
We used the updated codebook to evaluate all policies obtained in 2008 and 2011. Agreement between coders in assessing policy strength was very high, based on a 17% sample of randomly selected policies: The average kappa was 0.930 (229 of 240; 95%) for 2008 and 0.920 (158 of 168; 94%) for 2011. The coding system yielded extremely robust data.
The data indicate that, as of 2011, policy strength varied considerably by CCOI area (see Table 1). The CCOI policy areas with the highest frequencies of stringent policies in 2011 were ghostwriting (81; 6%), meals (62; 49%), gifts (44; 35%), and P&T committees (43; 34%). The lowest frequencies of stringent policies were in consulting (4; 3%), honoraria (8; 6%), CME (16; 13%), and vendor access (21; 17%). The frequency of “no policy” in 2011 was greatest for ghostwriting (38; 30%), honoraria (37; 29%), and speakers’ bureaus (30; 24%).
Each school’s 12 policy scores were averaged to create a measure of overall strength, or “policy strength average” (PSA). In calculating the PSA, the 12 scores were weighted equally, as all were identified as essential by the IOM, AAMC, and ABIMF–IMAP. Figure 1 shows the distribution of PSA scores in our sample. The mean PSA score for 2011 was 1.67 (median 1.75).
We then grouped schools by their PSA scores into four quartiles. The bottom quartile of the sample had a mean PSA score of 0.88 (just below “permissive”). The middle two quartiles had PSA scores averaging, respectively, 1.58 (between “permissive” and “moderate”) and 1.87 (just below “moderate”). The top quartile had a mean PSA score of 2.30 (between “moderate” and “stringent”).
No statistically significant differences were observed between the scores of public and private institutions, nor was hospital ownership a significant variable. However, medical schools with greater NIH funding were more likely than those with less NIH funding to have more stringent policies for meals, P&T committees, travel, scholarships, and speakers’ bureaus (see Table 2). Greater NIH funding was also correlated with higher PSA scores (F[1, 121] = 7.428, P = .007). In FY 2010, the mean NIH funding of the top PSA quartile was $98,311,270 compared with $53,881,289 for the bottom PSA quartile.
We then compared these findings with the recoded 2008 data and found dramatic changes. From 2008 to 2011, the frequency of “no policy” declined 57%–90% in 11 of 12 CCOI areas, with an overall decrease of 70% (see Table 1). The frequency of stringent policies increased in all areas, with the greatest increases occurring in honoraria (385%), speakers’ bureaus (355%), travel (244%), and scholarships (215%), for an overall increase of 178%. Average policy area scores rose 83%–198% in all CCOI areas except for CME, whose scores increased 5%.
We also found substantial improvements in medical schools’ overall policy strength. The mean PSA score grew 106%, from 0.81 to 1.67 (median, 0.92 to 1.75; see Figure 1). The proportion of schools with PSA scores of 0.1 or less—signifying no policies other than following the standards for commercial support devised by the Accreditation Council for Continuing Medical Education (ACCME)22—fell from 29% (22 out of 77) in 2008 to less than 2% (2 out of 127) in 2011. Average PSA scores by quartile also showed improvements: The bottom quartile’s mean PSA increased 11-fold, from 0.08 to 0.88; the middle two quartiles rose from 0.58 and 1.17, respectively, to 1.58 and 1.87; and the top quartile’s average PSA increased from 1.71 to 2.30.
We also examined changes in the PSA scores of the 75 medical schools whose policies were included in both the 2008 and the 2011 studies. Ninety-two percent (69) of these institutions showed an increase in their PSA scores, and 8% (6) had no change.
Discussion and Conclusions
This analysis provides an overview of recent changes in CCOI policies at U.S. MD-granting medical schools. Our data indicate that, from 2008 to 2011, medical schools made significant progress toward the standards recommended by the ABIMF-IMAP, IOM, and AAMC. In 2008, “no policy” was the most prevalent finding in 10 CCOI areas; in 2011, in no area was this the case. Moreover, average policy area scores increased dramatically for all categories except CME.
Our data also indicate striking shifts in schools’ overall policy strength. The vast majority of schools had an increase in PSA, and the mean PSA score more than doubled. Those in the bottom quartile, especially, have made strides: The proportion of schools with no enacted policies (PSAs of 0.1 or less) dropped sharply—from more than a quarter of our sample in 2008 to less than 2% in 2011. These results indicate a broad and rapid transformation of the policy landscape.
Many possible explanations may account for these changes. Perhaps the consensus of the ABIMF–IMAP, IOM, and AAMC spurred schools to reexamine and improve their policies. Additionally, medical school leadership is often closely attuned to developments at other institutions, so peer pressure may have played a role.15,19,21,23,24 External pressures have also increased. The media have avidly pursued stories of academic physicians’ conflicted relationships with drug and device companies; the failure of medical schools to control the problem has emerged as a common theme.12,25–28 These media stories prompted Senator Charles Grassley to initiate congressional investigations taking aim at medical schools that failed to manage inappropriate physician–industry ties, or even profited from such ties.4,29–36 These developments may have moved medical schools to implement stronger policies as a safeguard against scandal.
How close do schools now come to meeting the recommendations set forth by the AAMC, IOM, and ABIMF-IMAP? The highest average policy scores were for meals (2.22) and gifts (2.12). This may reflect that meals and gifts from industry are relatively trivial and difficult to defend; furthermore, banning them can take on great symbolic importance in modeling behavior for students and residents. Additionally, eliminating industry gifts and meals has comparatively little impact on institutional budgets.
More impressive, stringent ghostwriting policies have become the norm, with 64% of medical schools meeting national standards. This is an encouraging finding: Ghostwriting was the most neglected area in 2008 (with a “no policy” rate of 75%). What might explain this dramatic turnaround? Bans on ghostwriting are likely uncontroversial and thus relatively easy for schools to enact. It is also possible that awareness has increased. In 2008 and 2009, two of us (D.J.R., S.C.) conducted fieldwork at over a dozen medical schools; at schools without ghostwriting policies, deans and other top administrators often assured us “that doesn’t go on here.”23 Since then, a number of prominent publications have documented ghostwriting’s prevalence in medical research.5,37–41 These events may have spurred schools to enact stringent policies.
These accomplishments aside, much room for improvement remains. Less-than-stringent policies prevail in 11 of 12 CCOI areas. Nearly a third of medical schools still have no ghostwriting policy, and a majority of schools have no policies or permissive policies for samples, CME, consulting, honoraria, and speakers’ bureaus. CME is the most poorly controlled area: 106 (83%) schools have permissive policies, and 5 (4%) have none. Only the remaining 16 (13%) go beyond the industry-friendly standards for commercial support devised by the ACCME. Schools may believe these standards are sufficient for controlling conflicts of interest. Alternately, they may fear the financial consequences of stringent policies, which either prohibit industry support for CME or so restrict it as to make CME an unattractive “investment” for companies. Whatever the reasons, few schools have fully enacted the stringent standards recommended by the AAMC, IOM, and ABIMF–IMAP.
Our findings reveal not a race to the top but a shift from the bottom toward the middle. While the proportion of PSAs under 1.5 (more permissive than moderate overall) has fallen dramatically—from 81% in 2008 to 24% in 2011—PSAs in the moderate range (1.5–2.5) quadrupled, from 18% to 72%. All the while, the proportion of schools with strong policies overall—PSAs greater than 2.5—barely increased, from 1% (1 out of 77) in 2008 to 4% (5 out of 127) in 2011.
Why this move to the middle? Schools may perceive advantages to being in the middle of the pack, a position less likely to attract attention. It may also be easier to secure faculty buy-in for “middling” policies. Further research should determine whether this “race to the middle” persists or if a movement towards stringent policies emerges in the future.
This study has several limitations. First, some institutions may address CCOIs through informal means rather than relying on policies; conversely, formal policies may be ignored or not enforced.42 Our study did not evaluate implementation or enforcement practices. Second, our data collection ended on September 30, 2011, so policies enacted after this date are not reflected in our findings. Third, it is possible that the changes we observed in medical schools’ policies partly reflect differences in data collection practices. The 2008 data were obtained from contacts with medical schools’ compliance officers, whereas in 2011 we relied primarily on online searches, as most schools had begun posting their policies on the Internet. However, the accuracy of our results is supported by similar findings, using different methods, by the American Medical Student Association (AMSA). AMSA’s PharmFree Scorecard has shown improvements since 2009 in the letter grades earned by medical schools based on five “domains” of CCOI.24,43 These similarities suggest that our study provides an accurate view of the changing state of medical schools’ policies.
Our findings point to the ongoing need for medical schools to address consistently and comprehensively the challenges presented by CCOIs. Medical schools have made significant and rapid progress, yet the work is far from done: Most schools still lag behind national standards in many areas. Wider adoption of stringent CCOI policies is crucial to eliminate undue industry influence in clinical care and preserve public trust in medicine.
To promote change, first, medical school deans, compliance officers, and faculty must work to enact stronger policies. Second, the AAMC should use its leverage as a sponsor of the Liaison Committee on Medical Education (LCME) to make adherence to the AAMC’s rigorous guidelines16 a condition of LCME accreditation. Third, the NIH and other public funding agencies must take advantage of their unique leverage to set stringent, uniform conflict-of-interest requirements for grant recipients. A new Department of Health and Human Services (HHS) rule44 is a step in the right direction. It requires institutions to develop management plans for any “significant financial interests” (SFIs) judged to pose a financial conflict of interest (FCOI). It also lowers the reporting threshold for SFIs, from $10,000 to $5,000. However, the HHS rule provides no firm standards for determining whether an SFI constitutes an FCOI or whether an FCOI should be prohibited.45
To encourage and facilitate further change, IMAP’s database of CCOI policies is publicly available, in a searchable format.46 The database currently houses policies from 133 medical schools and is updated regularly as we receive new information. Our hope is that deans, compliance officers, faculty, and students will use the database to examine, compare, and implement strong policies to manage this ongoing problem. Medicine must act swiftly and effectively to address the challenge of CCOIs so as to protect scientific integrity and patient well-being.
1. Wazana A. Physicians and the pharmaceutical industry: Is a gift ever just a gift? JAMA. 2000;283:373–380
2. Dana J, Loewenstein G. A social science perspective on gifts to physicians from industry. JAMA. 2003;290:252–255
3. Adair RF, Holmgren LR. Do drug samples influence resident prescribing behavior? A randomized trial. Am J Med. 2005;118:881–884
4. Campbell EG. Doctors and drug companies—scrutinizing influential relationships. N Engl J Med. 2007;357:1796–1797
5. McHenry LB, Jureidini JN. Industry-sponsored ghostwriting in clinical trial reporting: A case study. Account Res. 2008;15:152–167
7. Blumenthal D. Doctors and drug companies. N Engl J Med. 2004;351:1885–1890
8. Chimonas S, Rothman DJ. New federal guidelines for physician–pharmaceutical industry relations: The politics of policy formation. Health Aff (Millwood). 2005;24:949–960
9. Armstrong D. Lawsuit says Medtronic gave doctors array of perks. Wall St J. September 25, 2008:B1
10. Harris G, Carey B, Roberts J. Psychiatrists, troubled children and drug industry’s role. N Y Times. May 10, 2007:A1
11. Abelson R. Whistle-blower suit says device maker generously rewards doctors. N Y Times. January 24, 2010:C1
13. Petersen M. Undisclosed financial ties prompt reproval of doctor. N Y Times. August 3, 2003:A18
15. Brennan TA, Rothman DJ, Blank L, et al. Health industry practices that create conflicts of interest: A policy proposal for academic medical centers. JAMA. 2006;295:429–433
17. New York Times Editorial Board. . Seducing the medical profession. N Y Times. February 2, 2006:A22
18. Chimonas S, Patterson L, Raveis VH, Rothman DJ. Managing conflicts of interest in clinical care: A national survey of policies at U.S. medical schools. Acad Med. 2011;86:293–299
19. Rothman DJ, Chimonas S. New developments in managing physician–industry relationships. JAMA. 2008;300:1067–1069
21. Chimonas S, Patterson L, Rothman D. Toeing the line. With added scrutiny, vendors heeding physician marketing rules. Mod Healthc. 2008;38:20
23. Rothman DJ, Chimonas S. Academic medical centers’ conflict of interest policies. JAMA. 2010;304:2294–2295
27. Ramshaw E. University of Texas officials vow to strengthen ethics rules for researchers. Dallas Morning News. February 11, 2009 http://search.proquest.com/newsstand
. Accessed June 20, 2013
29. Chimonas S, Stahl F, Rothman DJ. Exposing conflict of interest in psychiatry: Does transparency matter? Int J Law Psychiatry. 2012;35:490–495
34. White G, Schneider C. NIH puts pressure on Emory over funds. Atlanta J Constitution. October 15, 2008:A1
35. Armstrong D. NIH freezes large grant to Emory. Wall St J. October 15, 2008:A8
36. Wadman M. Straight talk with…Charles Grassley. Nat Med. 2008;14:1006–1007
37. Lacasse JR, Leo J. Ghostwriting at elite academic medical centers in the United States. PLoS Med. 2010;7:e1000230
38. Collier R. Prevalence of ghostwriting spurs calls for transparency. CMAJ. 2009;181:E161–E162
40. Ross JS, Hill KP, Egilman DS, Krumholz HM. Guest authorship and ghostwriting in publications related to rofecoxib: A case study of industry documents from rofecoxib litigation. JAMA. 2008;299:1800–1812
41. Anekwe TD. Profits and plagiarism: The case of medical ghostwriting. Bioethics. 2010;24:267–272
42. Austad KE, Avorn J, Franklin JM, Kowal MK, Campbell EG, Kesselheim AS. Changing interactions between physician trainees and the pharmaceutical industry: A national survey. J Gen Intern Med. 2013;28:1064–1071
44. Department of Health and Human Services. . Responsibility of applicants for promoting objectivity in research for which public health service funding is sought and responsible prospective contractors. Final rule. Fed Regist. 2011;76:53256–53293