Conflict of Interest in Seminal Hepatitis C Virus and Cholesterol Management Guidelines

Key Points

Question  Are recent clinical guidelines adhering to Institute of Medicine (IOM) standards for limits on conflicts of interest (COI)?

Findings  In this document review of 2 guidelines, the IOM standards for limits on COI among guideline committees were met by the cholesterol guideline but not by the hepatitis C virus guideline. Standards requiring no COI among committee chairs and cochairs were not met by either guideline.

Meaning  Levels of COI among guideline committees remain an important problem, and adoption of consistent COI frameworks may help ensure that guidelines are developed in a transparent and trustworthy manner.

Importance  Little is known regarding whether Institute of Medicine (IOM) standards for managing conflicts of interest (COI) have been met in the development of recent important clinical guidelines.

Objective  To evaluate adherence to the IOM standards for limits on commercial COI, guideline development, and evaluation of evidence by the 2013 American College of Cardiology and American Heart Association cholesterol management guideline and the 2014 American Association for the Study of Liver Diseases and Infectious Diseases Society of America hepatitis C virus management guideline.

Design, Setting, and Participants  This study was a retrospective document review of the June 2014 print version of the cholesterol guideline and the final September 2015 print version of the hepatitis C virus guideline. Each guideline was assessed for adherence to the IOM standards for commercial COI published in the 2011 special report Clinical Practice Guidelines We Can Trust.

Main Outcomes and Measures  The IOM standards call for no commercial COI among guideline committee chairs and cochairs and for less than 50% of committee members to have commercial COI. Guideline and contemporaneous article disclosure statements were used to evaluate adherence to these standards. Each guideline was also reviewed for adherence to other IOM standards for guideline development and evidence review.

Results  Among the 16 cholesterol guideline committee members, 7 (44%) disclosed commercial COI, all 7 reported industry-sponsored research, and 6 (38%) also reported consultancy. Of 3 guideline chairs and cochairs, 1 (33%) disclosed commercial COI. Review of contemporaneous articles identified additional commercial COI. Among the 29 hepatitis C virus guideline committee members, 21 (72%) reported commercial COI. Eighteen (62%) disclosed industry-sponsored research, 10 (34%) served on advisory boards, 5 (17%) served on data safety monitoring boards, 3 (10%) were consultants, and 3 (10%) reported other honoraria. Of 6 guideline cochairs, 4 (67%) disclosed commercial COI. All 4 disclosed additional COI in other publications that were not listed in their guideline disclosures. Contemporaneous literature review revealed an additional cochair with commercial COI. Of the 9 IOM guideline development and evidence standards, the cholesterol guideline met 5 (56%), and the hepatitis C virus guideline met them all.

Conclusions and Relevance  Neither the cholesterol guideline nor the hepatitis C virus guideline fully met the IOM standards for commercial COI management, and discordance between committee leader guideline disclosures and those in contemporaneous articles was common. Adherence to additional IOM standards for guideline development and evidence review was mixed. Adoption of consistent COI frameworks across specialty societies may help ensure that clinical guidelines are developed in a transparent and trustworthy manner.

Guidelines provide an authoritative synthesis of evidence and clinical experience that often proves influential in clinical care and insurance coverage decisions.^1-3 In 2011, responding to concerns regarding possible bias among guideline committee members, the Institute of Medicine (IOM) published standards limiting the types and magnitude of allowable conflicts of interest (COI) as part of a broader set of guideline development recommendations.⁴

Little is known about whether the IOM standards are being met in the development of recent guidelines.⁵ Although COI standards should be met by all guidelines, we selected for evaluation 2 recent guidelines that may prove to be among the most consequential for many years to come. The first was the 2013 cholesterol management guideline from the American College of Cardiology and American Heart Association (ACC/AHA),⁶ which moved away from the paradigm of treating to achieve specific target cholesterol levels. Second, starting in 2014, the American Association for the Study of Liver Diseases and Infectious Diseases Society of America (AASLD/IDSA) jointly developed guidelines for the use of novel drug treatments for chronic hepatitis C virus (HCV) infection.⁷ Both guidelines provide recommendations that help determine the use of medications to treat conditions that affect millions of Americans, with enormous clinical and economic implications.^8-12

In this retrospective document review, we compare the COI of members of the 2 committees that developed these significant guidelines. We use publicly available material to assess whether each committee adhered to IOM COI standards. We also examine the concordance of disclosures of committee leaders with the contemporaneous disclosures they provided in peer-reviewed literature. Last, we evaluate if each guideline met other IOM standards for development and evidence evaluation. Our objective is to provide information with which to evaluate whether the development of these 2 influential guidelines met standards to support broad acceptance of their rigor and objectivity.

Institutional review board approval was not sought for this article because it was a document review. The 2013 ACC/AHA cholesterol guideline (“cholesterol guideline”) was developed between 2008 and 2012 (months not specified). It was published electronically in November 2013 and then in print in June 2014.⁶ We used the COI disclosure information provided in the print version of the guideline for our analyses.

The development of the AASLD/IDSA HCV treatment guideline (“HCV guideline”) began in October 2013. The guideline was initially disseminated on the web (http://hcvguidance.org⁷) beginning in January 2014 before the first print version was published in September 2015. The formatting of COI disclosure statements for the committee members changed while the guideline was available online. We used the final version of disclosure information included in the 2015 print version of the guideline.

The IOM categorizes COI as intellectual or financial.¹³ Financial COI are subdivided into commercial or noncommercial types. Commercial COI include activities such as clinical services for which income is paid, consulting, board membership for which compensation is received, service as an expert witness, industry-sponsored research, patents, royalties, stock ownership, or other financial interests (Figure). Noncommercial COI include research grants and other types of support from entities, such as government, foundations, or nonprofits.^4,14 We focused solely on commercial COI, which have been considered to represent the greatest threat to external trustworthiness of the resulting guidelines and, as a result, have received the greatest emphasis in disclosure policies.¹³

We assessed each guideline for adherence to the IOM standards for commercial COI published in the 2011 special report Clinical Practice Guidelines We Can Trust.⁴ The IOM standards require that (1) less than 50% of the committee members have any commercial COI and (2) all chairs and cochairs have no commercial COI.⁴ The time span for COI disclosure differed between the 2 guideline committees. For the cholesterol guideline, the disclosure period covered all of guideline development from 2013 to 12 months before committee initiation in 2008 (month not indicated). For the HCV guideline, the disclosure period covered the time from publication in 2015 to the start of committee meetings in October 2013.

We also compared the commercial COI disclosures of each committee chair and cochair with their contemporaneous COI disclosures in relevant published articles in the peer-reviewed literature. For this analysis of COI concordance, we searched PubMed for each committee chair and cochair as authors of other publications during the period of guideline development (December 2008 to December 2012 for the cholesterol guideline and October 2013 to May 2015 for the HCV guideline). We included only publications directly related to cholesterol or HCV management in which chairs and cochairs were first or last author.

The IOM also provided guidance that addressed other elements, described in the foreward as “essential to developing sound practice guidelines.”⁴ Our analysis included standards relevant to guideline development and evidence review that were objective and able to be measured empirically. Each guideline’s text was evaluated for language indicating adherence to the standards.

Table 1 summarizes the findings across both guideline committees. The cholesterol guideline committee consisted of 16 members, with 1 chair and 2 cochairs. The disclosure standards were established by the ACC/AHA⁶ and required disclosure of all COI starting 1 year before committee initiation in 2008 through the end of the guideline development process in 2013. The cholesterol guideline development was initially managed by the National Heart, Lung, and Blood Institute (NHLBI) and then transferred to the ACC/AHA in 2013. The NHLBI and ACC/AHA had similar COI disclosure requirements: both required chairs to be without COI.⁶ The ACC/AHA also specified that a minority of committee members have COI. The cholesterol guideline development was funded by the ACC and AHA. Industry funding was not directly accepted for guideline development.¹⁵

Of the 16 committee members, 7 (44%) disclosed 1 or more types of commercial COI. These 7 individuals had industry-sponsored research grants, and 6 (38%) were also consultants.

The committee chair disclosed no commercial COI from 12 months before the start of guideline development in 2008 through guideline publication in 2013 (months not specified). However, in contemporaneous published articles, the chair disclosed service as a consultant and honoraria for educational activities from manufacturers of cholesterol treatments. At the time he was interviewed in 2013 about his relationships with drugmakers, he said that when asked to chair the committee, he disclosed to the NHLBI honoraria and consultancies with 6 companies, all of which made drugs to treat hyperlipidemia.^16,17 He indicated he had been assured that, if he had no such financial relationships for 1 year before serving on the committee, he would be eligible to be chair. Therefore, he ended relationships in May 2008 with the 6 companies and, in March 2009 at the first meeting of the cholesterol guideline committee, disclosed that he had no commercial COI.¹⁶ We cannot ascertain why his later published article¹⁸ still contained a listing of these former commercial relationships. It may be that he submitted his disclosure for the article before ending these relationships or that the disclosure time frame requested by the journal extended back to a time earlier than May 2008.

One cholesterol committee cochair disclosed commercial COI (only research grants from industry) as part of the guideline disclosure process. In contemporaneous publications, this individual also disclosed service as a consultant, advisory board membership, honoraria for educational activities, speaker honoraria, unspecified honoraria, and additional industry-sponsored research grants from companies that make cholesterol treatments.

Of the 9 additional IOM guideline development and evidence standards, the cholesterol guideline met 5 (56%). Most failures to meet standards were related to the narrow scope of the types of evidence eligible for consideration within the review provided to the committee (Table 2).

The HCV guideline committee consisted of 29 members, with 6 cochairs and no designated chair. The disclosure standards were established by the AASLD/IDSA and required disclosure of all COI during guideline development starting in October 2013. Conflict of interest statements for all committee members were updated shortly before guideline publication.⁷ Hepatitis C virus guideline development was funded by the AASLD and IDSA. The International Antiviral Society–USA provided administrative support. Industry funding was not directly accepted for guideline development.⁷

Of the 29 committee members, 21 (72%) reported 1 or more types of commercial COI. Eighteen (62%) received industry-sponsored research grants, 10 (34%) served on advisory boards, 5 (17%) served on data safety monitoring boards, 3 (10%) were consultants, and 3 (10%) had other unspecified honoraria (Table 1).

Of the 6 HCV guideline committee cochairs, 4 (67%) disclosed commercial COI. Two (50%) received industry-sponsored research grants, 1 (25%) served on a data safety monitoring board, and 1 (25%) had other unspecified honoraria. Disclosures in contemporaneous articles indicated that 1 additional cochair had commercial COI (industry-sponsored research), and there were several other important differences found for the other cochairs. For example, one cochair, who disclosed only industry-sponsored research in the guideline, indicated in peer-reviewed articles service as a consultant, payment for clinical trials, unspecified personal fees, and additional industry-sponsored research grants that were not included in the guideline disclosure statement. For all 4 cochairs who disclosed commercial COI for the guideline, there were additional commercial COI disclosed in contemporaneous articles.

The HCV guideline met all additional IOM guideline development and evidence standards. These results are summarized in Table 2.

Our analysis suggests that management of levels of commercial COI among guideline committees remains an important problem 5 years after the IOM standards were published. We found that the ACC/AHA cholesterol guideline adhered to the first IOM COI standard, but one of its committee leaders disclosed commercial COI, whereas the chair later acknowledged ending commercial links to relevant drugmakers immediately before assuming his chairperson role. The AASLD/IDSA HCV guideline did not come close to meeting either IOM COI standard, with almost three-fourths (21 of 29) of all committee members and two-thirds (4 of 6) of its cochairs disclosing commercial COI. Concerns about COI are compounded by our finding that contemporaneous COI disclosures in published articles by guideline committee chairs and cochairs frequently did not match, with additional and different relationships listed than those in their guideline COI disclosures.

The recommendations of the cholesterol and HCV guidelines evaluated herein differ from contemporaneous guidelines in a way that adds to questions about the possible influence of commercial COI. In 2014, cholesterol management guidelines were jointly developed by the Department of Veterans Affairs (VA) and the US Department of Defense (DoD).¹⁹ No members of this guideline group declared commercial COI. The VA/DoD guideline recommended a more conservative approach to cholesterol management through the use of higher thresholds for initiating statin therapy and recommendation of more moderate dosing levels.

Another VA expert committee without COI also produced HCV management guidelines that differed in important ways from that developed by the AASLD/IDSA. The 2014 VA HCV guideline had several recommendations for less expensive first-line treatment options and differed on the critical question of whether to consider postponing treatment for some patients.²⁰ The VA guideline gave cautious credence to this approach, whereas the AASLD/IDSA guideline did not.

There are several limitations to our study. First, we analyzed only 2 guidelines, and our findings may not be representative. We thought it was useful to focus our analysis on guidelines recommending major changes in the paradigm of appropriate care for large patient populations. Second, the results of our concordance analysis must be interpreted cautiously. We could not assess whether differing disclosure statements between guidelines and published articles reflected the severing of active commercial ties with drugmakers “just in time” in order to serve on guideline committees. Although not explicitly addressed in the IOM standards, it is unclear whether this practice reduces the chance of bias due to COI of guideline committee members, especially chairs and cochairs. Third, we could not evaluate with great precision whether the time spans for disclosure between the guidelines and contemporaneous articles may have differed. For example, journals may have required disclosure of conflicts dating back far longer than the guideline committees. However, this explanation seems unlikely given that both committees asked for disclosure over a broad multiyear time span.

Based on our findings, we propose 4 ways to improve the development of COI standards and COI disclosure by guideline committees. First, all specialty societies should agree to use the IOM framework and definition of the types of COI. In particular, we believe that the IOM was correct in its judgment that industry-sponsored research should be consistently recognized as relevant commercial COI.

Second, for consistency across settings of COI disclosure, we recommend that all guideline committees adopt the standard form for declaration of COI instituted in 2009 by leading medical journal editors.²¹ Consistency in COI forms should encourage and facilitate concordance between COI disclosures.

Third, we propose that the IOM standards for the level of acceptable commercial COI be adopted and implemented by all guideline committees. We believe there is no a priori reason why high-quality guidelines cannot be produced while following IOM requirements that a few committee members and none of the chairs and cochairs have commercial COI. Specialty societies have options for how to meet these standards. Some recruit a guideline writing committee entirely free of COI that meets with a second group composed of clinical experts with a rate of COI exceeding the 50% standard.^22,23 Before 2011, the ACC and AASLD had already adopted COI standards consistent with the IOM standards but, per our analysis, did not consistently adhere to their own policies.^24,25

Fourth, we propose that journal editors require that guideline committees meet the IOM standards in order for the resulting guideline to merit peer review and publication. Similarly, the National Guideline Clearinghouse could require adherence to the IOM standards as a prerequisite for inclusion. To improve consistency across disclosures, the National Institutes of Health, or a similar entity, could house a central COI disclosure database. Authors could be required to update their data annually, and journals could extract disclosures directly, avoiding the problem of inconsistent disclosures across journals.

We believe that our study highlights the need for broader and more explicit adoption of the IOM’s framework for COI management and a commitment from specialty societies to adhere to the IOM standards. Adoption of consistent COI frameworks will help ensure that the clinical guidelines establishing appropriate care in the United States are developed in a way that can merit the trust of patients, clinicians, and the broader public.

Corresponding Author: Steven D. Pearson, MD, MSc, Clinical Center, Department of Bioethics, National Institutes of Health, 10 Center Dr, Room 1C118, Bethesda, MD 20892 (pearsonsd@cc.nih.gov).

Accepted for Publication: November 8, 2016.

Published Online: January 17, 2017. doi:10.1001/jamainternmed.2016.8439

Author Contributions: Drs Jefferson and Pearson had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Both authors.

Acquisition, analysis, or interpretation of data: Jefferson.

Drafting of the manuscript: Both authors.

Critical revision of the manuscript for important intellectual content: Both authors.

Statistical analysis: Jefferson.

Administrative, technical, or material support: Both authors.

Study supervision: Pearson.

Conflict of Interest Disclosures: Dr Pearson reported receiving research grants during the past 2 years from the Blue Shield of California Foundation, the California Health Care Foundation, the Laura and John Arnold Foundation, and the National Pharmaceutical Council and reported membership dues paid to the Institute for Clinical and Economic Review during the past 2 years from Aetna, America’s Health Insurance Plans, Anthem, Blue Shield of California, CVS Caremark, Express Scripts, Harvard Pilgrim Health Care, OmedaRx, United Healthcare, Kaiser Permanente, Premera Blue Cross, Prime Therapeutics, Health Care Service Corporation, AstraZeneca, Genentech, GlaxoSmithKline, Johnson & Johnson, Merck, the National Pharmaceutical Council, Takeda Pharmaceuticals, Pfizer, Novartis, and Eli Lilly and Company. No other disclosures were reported.

Funding/Support: This study was supported by an intramural research grant from the National Institutes of Health.

Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The views expressed are the authors’ own and do not represent the views or policies of the National Institutes of Health, the US Department of Health and Human Services, or the US government.