Clinical research that involves living humans as subjects is the primary mechanism for translating the results of basic science into health-related products and services.1 Institutions in which clinical research is conducted are bound by Principle 7 of the Nuremberg Code, government regulations, assurances, and guidances to make proper preparations and provide adequate facilities to protect the persons involved in clinical research.2 Institutional review boards (IRBs) are the organizational bodies whose primary mission is to protect human subjects by reviewing clinical research protocols before implementation, approving changes in ongoing research protocols, and responding to reports of adverse events.3
In recent years, numerous reports have criticized IRBs for failing to adequately protect human subjects. The Inspector General of the U.S. Department of Health and Human Services concluded that many IRBs have insufficient research expertise and inadequate resources.4 The former U.S. Secretary of Health and Human Services5 and the former editor of the New England Journal of Medicine 6 suggested that financial conflicts of interest involving the pharmaceutical industry have undermined the integrity of the clinical research enterprise. A recent study found that institutions have failed to give adequate attention to issues of race and ethnicity in the conduct of clinical research.7 These criticisms, among others, have been leveled at some of the most research-intensive academic health centers in United States.5,8,9
In academic health centers (defined as medical schools and their affiliated and owned clinical facilities), faculty members constitute half of all IRB members.3 They are the primary source of research expertise for the IRB, and they serve as opinion leaders by informing the IRB's decisions and actions.10 Therefore, to address the criticisms noted above and given their major role, we sought to provide the first national data on the personal, professional, and research characteristics of faculty members who serve on IRBs.
The sample was drawn from the faculty roster database of the Association of American Medical Colleges (AAMC) for 1999. This database, which is updated annually, consists of detailed information regarding the personal and professional characteristics and activities of the 90,358 faculty members at 121 four-year U.S. medical schools (excluding those in Puerto Rico). We excluded 3,535 faculty members who did not perform direct patient care or conduct research; 1,406 with a rank other than instructor, assistant professor, associate professor, or full professor; 506 with primary affiliations in ancillary departments (administration, operations, library); and 674 with missing age data, leaving an eligible population of 84,237 faculty members. From this group, we drew a stratified sample of 5,084, based on respondents' professional age and primary professional activity, as indicated in the AAMC database. This sample was drawn to ensure adequate representation of young faculty, senior faculty, clinical faculty (those whose primary activity was patient care), and research faculty (those whose primary professional activity was research).
Our study was approved by the IRB at the Massachusetts General Hospital and the University of Massachusetts at Boston. The survey instrument, a modified version of a questionnaire we used in earlier work,11,12 was administered by mail between October 2001 and March 2002 by the Center for Survey Research at the University of Massachusetts. The initial mailing consisted of a survey questionnaire, a postage-paid return envelope, a fact sheet describing the characteristics of the study, and a letter from the principal investigator (DB). Nonrespondents to the first mailing were sent a reminder postcard and a second questionnaire mailing. Persistent nonrespondents were contacted by telephone and encouraged to participate.
These procedures resulted in 390 faculty members whom we determined to be ineligible because they no longer held a full-time faculty appointment in the medical school, were deceased, retired, out of the country, no longer located at the sampled institution, or unable to be located. Thus, we had a total of 4,694 eligible faculty members.
The primary dependent variable for our study was service on an IRB. On the questionnaire we asked, “In the last three years, have you been a member of an IRB (Institutional Review Board)?” The response categories were “yes” and “no.”
The independent variables were divided into the personal, professional, and research characteristics of the respondents. The personal characteristics included gender (male, female) and race/ethnicity (white non-Hispanic, black non-Hispanic, Asian/Pacific Islander, Native American or Alaskan Native, Hispanic/Latino). The race/ethnicity data were collapsed into a three-level variable representing white non-Hispanic, Asian, and underrepresented minority (URM) (black non-Hispanic, Hispanic, and Native American). Professional characteristics included academic rank (full professor, associate professor, assistant professor, and other ranks) and physician status (medical doctor [MD]/nonmedical doctor [non-MD]). Research characteristics included whether respondents had conducted any research in the three years preceding the study and if they conducted clinical research (defined as research involving living humans as research subjects). In addition, respondents were asked “Have you been a consultant to industry in the last three years?” Those who answered “yes” were considered to have been industrial consultants. Finally, our measure of research productivity was the number of articles published in the three years before our study in peer-reviewed journals.
We analyzed the data using standard statistical procedures. All analyses incorporated sampling weights to adjust for unequal probability of selection and nonresponse within the sampling strata. Differences in proportions were tested using chi-square and logistic regression analyses. Differences in means were tested using multivariate linear modeling procedures. All analyses were conducted using a statistical package that correctly computes the standard errors when determining statistical significance for survey data derived from complex sampling methods.
Characteristics of Respondents
Of the 4,694 eligible faculty members, 2,989 responded (66.5%). Table 1 shows the personal and professional characteristics of our respondents. Overall, 30.7% were female, 84.3% were white, non-Hispanic, and 75.5% were medical doctors. In terms of academic rank, 22.4% were full professors, 25.2% associate professors, 43.8% assistant professors, and 8.5% other academic ranks that included instructor and lecturers.
For comparison, Table 1 provides a breakdown of the universe of U.S. medical school faculty members from the AAMC Data Book for 1999. On only one variable (MD status) did our respondents differ by more than five percentage points from the AAMC's data. Given that the universe of faculty is extremely large and the fact that our sample and our respondents were not independent of that universe, statistical testing for differences between our respondents and the universe of medical school faculty was not needed or warranted.
We compared respondents to nonrespondents on two variables (gender and age). In terms of gender, 29.5% of nonrespondents were female, compared to 30.7% of respondents (p = .578). The mean age of nonrespondents was 56.8 years, compared to 60.3 for respondents (p = .357) (Data are not shown in Table 1).
Eleven percent of respondents (320 faculty members) reported that they had served on an IRB in the three years before our study (1999–2001) (Table 2).
Factors Predicting IRB Participation: Multivariate Results
Table 3 shows the results of a multivariate logistic regression analyses that modeled IRB participation. URM faculty members were 3.2 times more likely than white faculty members to serve on an IRB. At the same time, clinical researchers were 1.6 times more likely to be on an IRB than were faculty members who conducted nonclinical research. Assistant professors were significantly less likely than others to serve on IRBs (OR = 0.38, CI = 0.23 − 0.61).
IRB Service and Publication Rates
Overall, we found no significant difference in the average number of articles published in the three years before our study comparing IRB faculty members to non-IRB faculty members after statistically controlling for effects of gender, academic rank, race, physician status, and clinical research status (11.5 versus 10.1, p = .27). In stratified analyses by academic rank, we found no differences in numbers of publications comparing faculty IRB members versus non-IRB faculty members among full professors (17.6 versus 15.4, p = .38), associate professors (9.2 versus 7.2, p = .05), assistant professors (6.2 versus 6.7, p = .40) and other faculty ranks (4.6 versus 5.6, p = .40). In stratified analyses by race/ethnicity, there were no statistically significant differences for white faculty members (10.9 for IRB faculty versus 9.3, for non-IRB faculty, p = .25) and Asian faculty members (9.4 versus 9.5, p = .97). However, URM faculty members who served on an IRB published an average of eight articles in peer-reviewed journals in the three years before our study compared to 3.5 for URM faculty members who had not served on an IRB (p < .001), adjusting for all the other variables mentioned above (Table 4).
Our study provides the first U.S. portrait of the personal, professional, and research characteristics of faculty members who serve on IRBs in the nation's medical schools. Overall, recent service on an IRB (in the three years before our study) was limited to a small group of medical school faculty members (11%) in 2001. This finding may reflect a lack of recognition for this activity in the reward structure of medical schools that tend to focus primarily on research accomplishments, such as published articles and grants, rather than research-related activities, such as serving on the IRB.13 It may also be that there are other forms of institutional service that compete for the attention of faculty members. However, it is possible that this amount of faculty participation is sufficient to meet the needs of IRBs.
Several characteristics of faculty members are shown to be related to IRB participation. The strongest predictor of IRB service is race/ethnicity, given that URM faculty members are clearly overrepresented on IRBs relative to their rates in the faculty population. This finding may be the result of greater pressure on URM faculty members to serve on IRBs, an explanation consistent with previous research demonstrating that black and Hispanic faculty members felt pressure to serve on committees because of their race and ethnicity.14 An alternative explanation is that URM faculty members seek out IRB service at rates higher than non-URM faculty members. Regardless of the reason, the higher rates of participation on IRBs may signal a positive role for URM faculty members on matters related to clinical research management, and may lead to greater attention paid to issues of race/ethnicity in the design and conduct of culturally sensitive clinical trials. Additional research is needed to understand better the differential rates of participation based on race/ethnicity and the effect, if any, of such participation on the conduct of research and the performance of IRBs.
Another characteristic that predicts faculty IRB participation is involvement in research generally and clinical research particularly. This finding suggests that faculty IRB members have personal research experience and knowledge that may inform their IRB-related activities. At the same time, clinical researchers serving on IRBs may feel pressure not to impede studies important to their individual areas of research, their departments, or to their colleagues—representing a potential nonfinancial conflict of interest. Alternatively, some faculty IRB members who conduct clinical research may intentionally block studies of other investigators when the proposed research might compete with their own studies. Future research should explore the effect, if any, of members' personal research experiences on their IRB-related activities.
Consulting relationships with industry are common among faculty IRB members. As noted earlier, we found that about half (47%) of all faculty IRB members had served as consultants to industry in the three years before our study. Our previous research among life science faculty members has shown that associations with industry are related to scientific behavior, including delays in publishing research and trade secrecy.15,16 It is possible that relationships with companies could affect members' IRB-related activities and attitudes as well. This finding may be a cause for some concern given that faculty members serve as leaders on the IRB. The fact that almost half serve as consultants to industry (most of whom are likely compensated for their services) certainly raises the issue of financial conflicts of interest. However, relationships with companies among faculty IRB members may have benefits, because presumably some direct experience with industry may facilitate a deeper understanding of the types of human subjects concerns that may arise in industry-funded studies. Additional research should address the nature, extent, and consequences (both positive and negative) of industry relationships among faculty IRB members.
Finally, our data suggest that serving on an IRB is not associated with decreased levels of research productivity as measured by publications. It may be that the demands of IRBs are not sufficient to reduce research activities or that IRB service provides members with research ideas and insights that bolster their research productivity. Alternatively, faculty IRB members may devote additional time and energy to research to make up for their investment in IRB-related activities, or they may be more efficient than non-IRB faculty members. It may also be that faculty members with high rates of publication self-select as members of the IRB. Regardless, concerns that IRB service detracts from research productivity were not supported by our data and, if confirmed by further research, this information could be used to recruit additional researchers to serve on the IRB.
Our study had a number of limitations. First, because we did not study all IRB members, our results only apply to IRB members who hold faculty appointments and are not applicable to IRB members who do not hold faculty appointments, such as allied health professionals and noninstitutional IRB members. Second, our results may not apply to nonmedical school faculty such as those in schools of public health and the allied health sciences who may also serve on medical school IRBs. Third, the cross-sectional nature of our study precludes making causal assertions. Fourth, our results do not apply to faculty in medical schools that use IRBs not affiliated with the medical school. Fifth, we did not ask about other forms of academic–industry relationships such as equity, bonuses, and ownership of businesses. Finally, our study did not directly address the IRB-related activities of faculty such as the amount of time devoted to IRB activities.
Despite these limitations, our study provides the first national portrait of the personal, professional, and research characteristics of faculty members who serve on IRBs. Our findings have practice and policy implications for recruiting faculty members to serve on IRBs, addressing the racial, ethnic, and gender representation of faculty IRB participants and addressing issues of potential conflicts of interest among those who shoulder a portion of the responsibility for protecting human research subjects.
1. Commonwealth Fund Task Force on Academic Health Centers. From Bench to Bedside: Preserving the Research Mission of Academic Health Centers. New York: The Commonwealth Fund, 1999.
2. Institute of Medicine. Preserving Public Trust: Accreditation and Human Research Participant Protection Programs. Washington, DC: National Academy Press, 2001.
3. National Institutes of Health. Final Report: Evaluation of the NIH Implementation of Section 491 of the Public Health Service Act, Mandating a Program of Protection for Research Subjects. Arlington, VA: National Institutes of Health, 1998.
4. Christian MC, Goldberg JL, Killen J, et al. A central institutional review board for multi-institutional trials. N Engl J Med. 2002;346:1405–8.
5. Shalala D. Protecting research subjects: what must be done. N Engl J Med. 2000;343:808–10.
6. Angell M. Is academic medicine for sale? N Engl J Med. 2000;342:1516–8.
7. Gifford AL, Cunningham WE, Heslin KC, et al. Participation in research and access to experimental treatments by HIV-infected patients. N Engl J Med. 2002;346:1373–82.
8. Kaplan S, Brownlee S. Dukes hazards: did medical experiments put patients needlessly at risk? US News & World Report, 1999 May 24;66–70.
10. Lemmens T, Freedman B. Ethics review for sale? Conflict of interest and commercial research review boards. Milbank Q. 2000;78:547–84, iii-iv.
11. Campbell EG, Weissman JS, Causino N, Blumenthal D. Data-withholding in academic medicine: characteristics of faculty denied access to research results and biomaterials. Research Policy. 2000;29:303–12.
12. Blumenthal D, Causino N, Campbell EG, Weissman JS. The relationship of market forces to the satisfaction of faculty at academic health centers. Am J Med. 2001;111:333–40.
13. Ludmerer KM. Time to Heal: American Medical Education from the Turn of the Century to the Era of Managed Care. Oxford and New York: Oxford University Press, 1999.
14. Palepu A, Carr PL, Friedman RH, Amos H, Ash AS, Moskowitz MA. Minority faculty and academic rank in medicine. JAMA. 1998;280:767–71.
15. Blumenthal D, Campbell EG, Causino N, Louis KS. Participation of life-science faculty in research relationships with industry. N Engl J Med. 1996;335:1734–9.
16. Blumenthal D, Campbell EG, Anderson MS, Causino N, Louis KS. Withholding research results in academic life science. Evidence from a national survey of faculty. JAMA. 1997;277:1224–8.