Abstract
This cross-sectional study evaluates the association of appearance of conflicts of interest with voting behavior at FDA advisory committee meetings.
To ensure objectivity on the 49 advisory committees (ACs) of the US Food and Drug Administration (FDA), the FDA reviews disclosure statements submitted by members for 2 potentially disqualifying types of financial conflict of interest (COI): “section 208 conflicts” (regulated under 18 USC §208) and, for those without section 208 conflicts, “section 502 conflicts” (regulated under 5 CFR §2635.502). While section 208 conflicts are created by current financial interests, section 502 conflicts are typically created by past financial interests or personal and business relationships that could create the appearance of a conflict. For both, the FDA may either exclude the individual from participation or allow participation when permitted by statute. The FDA publicly discloses section 208 conflicts but not section 502 conflicts and has recently requested comments about whether these conflicts should be publicly disclosed.
We examined the prevalence of section 502 conflicts and their associations with voting behavior at FDA AC meetings.
Methods
We obtained internal agency information for AC meetings with votes on drugs and devices held between 2008 and 2014 for which COI screening was required. For meetings with section 502 conflicts present, we determined the prevalence and characteristics of both section 208 and section 502 conflicts and, for meetings related to particular products, assessed whether members with section 502 conflicts were more likely to vote with the majority than were members without these conflicts. For meetings with no section 208 conflicts and where the vote was not unanimous, we assessed whether the prevalence of section 502 conflicts was associated with the percentage of votes favorable to the product and with the meeting outcome. Finally, we examined whether excluding members with section 502 conflicts would have changed the meeting outcome. Institutional review board approval was not required because the study was not considered human subjects research.
Results
Of 385 AC meetings that met entry criteria, 27.3% (n = 105) included at least 1 voting member with a section 502 conflict; there was no trend over time. Of 1482 voting members in these meetings, 12.6% (n = 187) had section 502 conflicts only; 0.8% (n = 12) had section 208 conflicts only, and 0.9% (n = 14) had both. More detail is reported in Table 1. Individuals with only section 502 conflicts were typically organizational officers (36.9%; n = 69), had conflicts with sponsoring firms (89.3%; n = 167), and had imputed COIs (ie, COIs assigned based on members’ personal relationships or employment) (61.0%; n = 114). Voting members with only section 208 conflicts were more likely than those with 502 conflicts to have conflicts with competing firms (58.3% [n = 7] vs 8.6% [n = 16]; Pearson χ2 P < .001) but were otherwise similar. The median values of section 502 and 208 conflicts were $30 000 and under $50 000, respectively (the financial value of section 208 conflicts is reported in ranges).
Table 1. Characteristics of Voting Members at All Meetingsa With at Least 1 Section 502 Conflict, 2008-2014.
| Characteristic | Members With Conflicts, No. (%) | P Value | |||
|---|---|---|---|---|---|
| Section 502 Only (n = 187) |
Section 208 Only (n = 12) |
Both (n = 14) |
Neither (n = 1269) |
||
| Committee member status | .73b | ||||
| Standing | 75 (40.1) | 3 (25.0) | 5 (35.7) | 513 (40.4) | |
| Temporary | 112 (59.9) | 9 (75.0) | 9 (64.3) | 756 (59.6) | |
| Participant’s role in the affected organization | .56c | ||||
| Principal investigator | 39 (20.9) | 3 (25.0) | NA | NA | |
| Officerd | 69 (36.9) | 4 (33.3) | NA | NA | |
| Consultant-speaker | 44 (23.5) | 1 (8.3) | NA | NA | |
| Employee | 14 (7.5) | 1 (8.3) | NA | NA | |
| Others | 21 (11.2) | 0 | NA | NA | |
| Stakeholders | 0 | 3 (25.0) | NA | NA | |
| Source of conflict | <.001c | ||||
| Sponsor | 167 (89.3) | 5 (41.7) | NA | NA | |
| Competitor | 16 (8.6) | 7 (58.3) | NA | NA | |
| Both | 4 (2.1) | 0 | NA | NA | |
| Nature of conflict | .53c | ||||
| Imputede | 114 (61.0) | 8 (66.7) | NA | NA | |
| Personal | 55 (29.4) | 4 (33.3) | NA | NA | |
| Both | 18 (9.6) | 0 | NA | NA | |
Abbreviation: NA, not applicable.
Includes product and nonproduct meetings at the Center for Drug Evaluation and Research and the Center for Devices and Radiological Health. Individual advisory committee members may have attended more than one meeting. The 502 conflicts are identified only if no section 208 conflict exists for a particular meeting. Individuals may have waivers for both conflicts if they are issued for different particular matters.
Pearson χ2 test, comparing all 4 categories.
Pearson χ2 test, comparing only section 502 conflict with only section 208 conflict.
A person elected or appointed to manage the organization’s daily operations, such as a president, vice president, secretary, treasurer, or board chair.
Conflicts assigned to an advisory committee member based on their personal relationships or employment (eg, a research contract with a potential members’ employer in which the member has no involvement).
Of 151 votes on specific medical products by members with only section 502 conflicts, 132 (87.4%) concurred with the majority (Table 2). We found no association between the prevalence of section 502 conflicts and voting outcomes. Excluding members with section 502 conflicts generally would have produced voting margins more favorable to the product, but in no instance would the meeting outcome have changed between favoring and opposing the product. In 1 instance, excluding 2 members with only a section 502 conflict would have led to a tie vote rather than a vote opposing the product.
Table 2. Voting Patterns Among Members in 86 Product Meetings With at Least 1 Section 502 Conflict, 2008-2014.
| Characteristic | Opposed to Majority | Concur With Majority | Abstain |
|---|---|---|---|
| Individual financial conflict status, No. (%)a | |||
| Only section 502 (n = 151) | 13 (8.6) | 132 (87.4) | 6 (4.0) |
| Only section 208 (n = 8) | 0 | 8 (100) | 0 |
| Both (n = 12) | 2 (16.7) | 10 (83.3) | 0 |
| Neither (n = 1015) | 121 (11.9) | 864 (85.1) | 30 (3.0) |
| Meeting-Level Analysisb |
Any Section 502 Conflict (n = 47) |
Section 502 Conflict With Sponsoring Firm Only (n = 42) |
NA |
| Those with section 502 conflicts vs those favoring the product, % | |||
| Spearman ρ | −0.07 | −0.19 | NA |
| P value | .66 | .22 | NA |
| Those with section 502 conflicts vs majority favoring the product (yes/no) | |||
| P value | .55c | >.99c | NA |
| Impact of exclusions of members with section 502 conflicts on vote margin, No. (%) | |||
| Voting margin less favorable | 16 (34.0) | 15 (35.7) | NA |
| Voting margin more favorable | 28 (59.5) | 24 (57.1) | NA |
| Voting margin unchanged | 3 (6.4) | 3 (7.1) | NA |
| Impact of exclusions of members with section 502 conflicts on meeting outcomes | 1 meeting with 2 members each with 1 section 502 conflict changed from unfavorable to a draw | 1 meeting with 2 members each with 1 section 502 conflict changed from unfavorable to a draw | NA |
Abbreviation: NA, not applicable.
Pearson χ2 test P = .70, comparing 4 categories of financial conflict status and 3 categories of voting outcome.
Excludes 39 product meetings with unanimous voting or where a participant had section 208 conflicts.
Wilcoxon rank sum test.
Discussion
Between 2008 and 2014, about a quarter of FDA AC meetings that met entry criteria included at least 1 participant with a section 502 conflict. Members with section 502 conflicts had voting patterns similar to those without such conflicts; there was no association between section 502 conflicts and meeting outcomes.
The study limitations were as follows: (1) including only meetings where at least 1 member had a section 502 conflict; (2) small sample sizes, particularly in the voting analyses; and (3) inability to control for other factors that might influence voting outcomes, such as member expertise and media coverage.
References
- 1.US Code. Acts affecting a personal financial interest. 18 USC §208.
- 2.US Code. Personal and business relationships. 5 CFR §2635.502.
- 3.US Food and Drug Administration Guidance for the public, FDA advisory committee members and FDA staff on procedures for determining conflict of interest and eligibility for participation in FDA advisory committees. https://www.fda.gov/downloads/RegulatoryInformation/Guidances/UCM125646.pdf. Accessed December 22, 2016.
- 4.US Food and Drug Administration Procedures for evaluating appearance issues and granting authorizations for participation in FDA advisory committees. https://www.fda.gov/downloads/RegulatoryInformation/Guidances/UCM508852.pdf. Accessed March 9, 2017.
- 5.US Food and Drug Administration FDA advisory committee members and 'appearance issues'. https://blogs.fda.gov/fdavoice/?s=ortwerth. Accessed December 22, 2016.