Routine Conflict of Interest Disclosure by Preclinical Lecturers and Medical Students' Attitudes Toward the Pharmaceutical and Device Industries FREE

To the Editor: Disclosure of potential conflicts of interest (COI) from relationships with the pharmaceutical and device industries is recommended¹ in publishing, research, and education of residents, faculty,² and students.³ Disclosure in education may foster critical evaluation of information and assessment for potential bias,³ but its effect is unclear.

In September 2010, Mount Sinai School of Medicine began mandating COI disclosure by lecturers to preclinical medical students. Using the natural experiment created by this new policy, we studied the association between routine COI disclosure with preclinical medical students and students' attitudes toward disclosure and industry interactions.

Policy implementation included a brief presentation to students introducing the disclosure policy, routine COI disclosure by faculty during lectures, and continuous student online access to faculty disclosures. We verified faculty posting of online disclosures but not disclosure during lectures.

We adapted a published survey⁴ to assess student attitudes toward COI disclosure and the appropriateness of industry gifts to physicians, industry-sponsored education, and industry-faculty relationships with responses on a 4-point Likert scale (1 = strongly disagree; 2 = disagree; 3 = agree; 4 = strongly agree). All year 1 (n = 143) and year 2 (n = 141) students were eligible to participate; results from the 2 classes were pooled.

We surveyed participants during class meetings at the beginning and end of the 2010-2011 academic year, comparing attitudes (combined agree and strongly agree responses) before and after policy implementation using a 2-sided Fisher exact test, with significance defined by P values of less than .05. The GraphPad 2005 software (QuickCalcs) was used for statistical analysis. Participation was voluntary and anonymous. The Mount Sinai institutional review board approved the study and exempted it from formal informed consent.

Among 188 lecturers to preclinical students, 146 (78.9%) posted disclosures; 130 (89.0%) reported no potential COI. Survey response rates at the beginning and end of the academic year were 66.2% (188/284) and 60.9% (170/279) and were similar in first- and second-year students; 5 first-year students left school. Predisclosure and postdisclosure samples did not differ with regard to student sex (52% male), age (predisclosure: 23.7 [SD, 1.91] years, postdisclosure: 24.3 [SD, 1.96] years), or class.

Nearly all students (>97.0%) favored disclosure in both surveys (Table 1). Attitudes toward academic-industry relationships changed after policy implementation (Table 2). Agreement increased over whether schools should limit industry meetings with students (104/184 [56.5%; 95% CI, 49.3%-63.5%] predisclosure vs 120/167 [71.9%; 95% CI, 64.6%-78.1%] postdisclosure; P = .004) and educators (74/186 [39.8%; 95% CI, 33.0%-47.0%] predisclosure vs 95/168 [56.5%; 95% CI, 49.0%-63.8%] postdisclosure; P = .002). Agreement decreased over whether industry should fund medical school programs (110/179 [61.5%; 95% CI, 54.2%-68.3%] predisclosure vs 72/163 [44.2%; 95% CI, 36.8%-51.8%] postdisclosure; P = .002). Attitudes did not change regarding the influence of COI on educators or educational content (Table 1). Few students believed that educational content or quality were influenced by educator relationships with industry.

There is little evidence documenting the effect of COI disclosure despite widespread acceptance of its importance.^1,5 We believe our study is the first to evaluate the association between COI disclosure and preclinical medical students' attitudes. Participants were more aware of the potential influence of industry relationships than students in past studies.⁶

Independent of the policy, many favored disclosure by educators and limiting industry interactions, and believed that industry interactions influence prescribing. Routine COI disclosure was associated with an increase in student desire for limitations in some industry relationships, but not with perceptions of disclosure, the effect of industry relationships on educational content, or instruction by faculty with relevant COI. The lack of association may be related to students' relationships to professors or the rarity of faculty COI. The introduction of the policy, rather than disclosure itself, may have influenced the students' attitudes.

There are several limitations, including the single-site, predisclosure and postdisclosure design, and potential social desirability bias. We are unaware of other changes over the year that might explain the results. We were unable to account for confounding (although there were no differences in responses by sex or class), or to verify disclosure accuracy or the presentation of disclosures during lectures.

Our findings suggest that a COI disclosure policy to students is feasible and may influence student attitudes toward industry prescribing but not education. Future studies should determine the generalizability and persistence of our findings.