Discrepancies in Conflict-of-Interest Disclosures Among Physicians Receiving Compensation for Monoclonal Antibody Drugs

Abstract

Background

Monoclonal antibody drugs are widely used, highly marketed, expensive compounds. Relationships between these drug manufacturers and physicians may increase the potential for bias in relevant studies.

Objective

The aim of this study is to determine the rate of disclosures among physicians receiving compensations for monoclonal antibody drugs (MAbDs).

Design

This is a retrospective, population-based, cross-sectional study.

Participants

The 50 physicians who received the highest financial compensation for selected MAbDs from 2016 to 2020 were included.

Main Measures

Payment data were obtained from the Open Payments Database, bibliometric data were obtained from SCOPUS, and disclosure data were obtained from relevant publications found in PubMed. The primary outcome was rate of disclosure concordance between self-declared conflict-of-interest and industry-reported payments documented in the Open Payments Database.

Key Results

Of the 50 physicians examined, 74% (N = 37) had publications examined. A cumulative 6170 payments totaling $18,484,228 were analyzed. A total of 418 relevant papers were reviewed. The rate of full disclosure (all relevant financial relationships disclosed) was 39.5%, partial disclosure (some but not all financial relationships disclosed) was 28.0%, and no disclosure was 26.3%. 6.2% did not require disclosure. Publications authored by dermatologists had the highest rate of full disclosure at 49.3%. There was no association between h-index and disclosure rate. Practice guidelines had the highest rate of full disclosure at 69.2% while basic science papers had the lowest (0%). Lastly, substantial variations in specific journal disclosure policies were found.

Conclusions

Substantial inconsistencies were found between self-reported disclosures and the Open Payments Database among physicians receiving high compensation for MAbDs. A policy of full disclosure for all publications should be adopted.

INTRODUCTION

While collaboration between industry and health care providers is important for advancing medical knowledge, industry compensation to physicians has the potential to create bias and influence clinical and research activities. Indeed, compensation to physicians is correlated with increased prescribing rates and formulary addition requests for the promoted drug.^{1, 2} Likewise, industry-funded studies have been associated with positive results supporting the funded product.^{3, 4} Thus, reporting industry relationships when publishing biomedical research is critical to enable public trust in the credibility of clinical recommendations and medical research.

Spending on medical marketing of drugs to consumers and physicians has been increasing and reached approximately $30 billion by 2016, with 66% of spending targeted to physicians.⁵ Much of the increase in spending can be attributed to competition for biologic therapies and drugs indicated for the treatment of dermatologic conditions, arthritis, neurological disorders, cardiac diseases, diabetes, and cancer.^{5, 6} The advent of biologic agents for the treatment of autoimmune and inflammatory disorders has drastically improved the treatment of chronic, disabling diseases. As such, the number of patients on biologic therapies has exploded; in one study population between 2007 and 2015, the proportion of patients on biologic therapy for Crohn’s and ulcerative colitis increased by 200% and 317%, respectively.⁷ Many biologic immunosuppressive therapies are approved for the treatment of several diseases; thus, industry compensation to physicians for these multiple-indication antibody drugs (MAbDs) creates conflicts of interest across multiple specialties.

The Physician Payment Sunshine Act was passed as part of the Patient Protection and Affordable Care Act in 2010.⁸ The purpose of this law is to increase transparency of financial relationships between physicians and teaching hospitals and pharmaceutical and medical device companies. The International Committee of Medical Journal Editors (ICMJE) requires physicians to disclose relevant conflicts of interest for publication in academic medical journals,⁹ yet multiple studies comparing author disclosure rates with industry-reported transactions available in the Open Payments Database have demonstrated broadly discordant disclosures.^10–12

Since spending on new expensive biologics heavily contributes to increases in industry compensation to physicians, we designed a study to examine the trends of disclosure of conflicts of interest in publications by physicians receiving compensation for multiple-indication biologic drugs. We hypothesized that we would identify discrepancies in payments to top-earning physicians by medical industry and self-disclosed conflicts of interest. We also sought to examine trends of self-disclosure by physician specialty and type of research study.

METHODS

This study was determined to be non-human subject research by the Weill Cornell Institutional Review Board (#22-11025386). Payment data were extracted from the Open Payments Database for companies that make payments for multiple-indication monoclonal antibody drugs. The Open Payments Database is a national program which collects and publishes information that is pertinent to financial relationships between drug and medical device companies and health care providers or teaching hospitals to provide the public with transparency in health care.¹³ Companies that produce, prepare, or compound a covered drug, biologic, or device and receive reimbursement from Medicare, Medicaid, or the Children’s Health Insurance Program are required to submit payment information to the Centers for Medicare and Medicaid Services, which is then shared with the public.¹³ The ICMJE requires authors, peer reviewers, editors, and editorial board members to disclose financial and non-financial relationships at time of submission, when asked to critique a manuscript, or when making final editorial decisions, respectively. The ICMJE defines a conflict of interest as any relationship or activity that may bias or be perceived as a bias that affects credibility of published articles. Financial relationships include but are not limited to consulting activities, speaking engagements, support for attending meetings, employment, research funding, ownership, gifts, and expert testimony. Non-financial relationships include personal relationships, rivalries, intellectual or financial competition, and unpaid leadership roles on boards, societies, or committees. The time frame for such disclosures is 36 months.⁹

Sample Selection

Drugs which were included in the analysis were monoclonal antibody drugs that had multiple FDA-approved indications for use. Eight drugs were included in the final analysis: Humira, Stelara, Entyvio, Cimzia, Tysabri, Remicade, Inflectra, and Renflexis. General payments made to physicians for these drugs were obtained from the Open Payments Database from years 2016 to 2020. General payments include consulting fees, non-publicly traded ownership, honoraria, gifts, compensation for food and beverage, travel and lodging expenses, entertainment, royalties or licensing agreements, speaking fees, and compensation for services other than consulting. Payments for food and beverage, travel, and yearly total payments under $750 were excluded from the analysis. Open Payments Database data were processed in R to identify the top physicians receiving compensation for services related to multiple-indication monoclonal antibody drugs. The top 50 paid physicians from any specialty were included in the analysis.

Publication Selection

Publications for each of the top 50 paid physicians were obtained from PubMed from years 2017 to 2021. Abstracts and keywords were screened for relevance. Only relevant articles were included in the analysis. A publication was deemed relevant if it related to the physiology or treatment of FDA-approved indications of a monoclonal antibody drug for which a physician received industry payment. For example, if Dr. X received payments for Remicade, and had two publications, one on the treatment of Crohn’s disease and one on medical student education, the former would be included in the analysis while the publication on medical student education would be excluded.

Bibliometric Data

The Scopus database was used to determine each physician h-index. H-index is a scholarly measure which determines the scientific outputs of researchers. It can be defined as “the number of papers (Np) papers have at least h citations each and the other (Np – h) papers have ≤h citations each.”¹⁴

Outcome Assessment

All authors were considered to have a potential conflict of interest based on the payments received and documented in the Open Payments Database. Each article’s disclosure statement was examined to determine if they adhered to ICMJE standards. Discrepancies in self-declared conflicts of interest were determined. Authors were considered to have “full disclosure reporting” if all relevant financial ties were disclosed, “partial disclosure reporting” if some but not all financial ties were disclosed, and “no disclosure reported” if no financial ties were disclosed. For example, if an author has received payments on behalf of both Stelara and Remicade and published a paper evaluating Remicade, ICMJE guidelines require that financial relationships to both must be disclosed. If a disclosure statement did not appear in the publication, a disclosure was deemed “no statement included” by the publishing body. For each paper which was examined, disclosure statements were crosschecked with payments from the previous calendar year.

Statistical Analysis

All data management was performed using R. Data analysis was performed using STATA. Descriptive statistics and chi-squared tests were performed to determine differences between groups.

RESULTS

In total, companies made 40,435 payments on behalf of these eight monoclonal antibody drugs, totaling $77,477,595 from 2016 to 2020. The 50 highest paid physicians accounted for 6170 (15%) of payments, totaling $18,484,228 (24%). Total payments and payments made to the top 50 paid physicians on behalf of these 8 monoclonal antibodies drugs are summarized in Table 1.

Table 1.

Summary of Payments Made to All Physicians and the Top 50 Paid Physicians on Behalf of the Eight Selected Monoclonal Antibody Drugs

Drug	Company	Total payment (USD)	Total number of payments (N)	Median payment (range)
Payments made to all physicians
Humira	AbbVie, Inc.	$37,303,024.28	18,167	$2000 (3.85–77,500)
Stelara	Janssen Pharmaceuticals	$1,3017,118.85	9006	$812.5 (3.82–154,500)
Tysabri	Biogen, Inc.	$11,715,917.33	4859	$2250 (3.54–15,187.5)
Entyvio	Takeda Pharmaceuticals	$7,875,559.32	2835	$2336.25 (0.91–125,000)
Cimzia	UCB, Inc.	$5,565,006.38	3237	$1150 (0.95–17,250)
Remicade	Janssen Pharmaceuticals	$1,309,560.12	1686	$99 (38–10,000)
Inflectra	Pfizer, Inc./Celltrion	$432,702.76	513	$12.7 (3.47–15,000)
Renflexis	Merck Sharp & Dohme Corporation	$258,706.75	132	$2065 (105–3500)
Payments made to top 50 paid physicians
Humira	AbbVie, Inc.	$7,702,896.7	2471	$3125 (3.85–77,500)
Stelara	Janssen Pharmaceuticals	$5,135,156.06	1923	$2575 (71.99–26,250)
Entyvio	Takeda Pharmaceuticals	$3,457,932.27	863	$3500 (2.77–26,250)
Cimzia	UCB, Inc.	$1,133,303.14	560	$1600 (1.32–15,117.39)
Tysabri	Biogen, Inc.	$578,636.83	171	$2765 (4.46–4500)
Remicade	Janssen Pharmaceuticals	$262,041.29	105	$3300 (44.95–4500)
Inflectra	Pfizer, Inc./Celltrion	$148,936.76	53	$2750 (9.16–10,000)
Renflexis	Merck Sharp & Dohme Corporation	$6,5325	24	$2625 (2065–3500)

Eighty-four percent (n = 42) of the 50 top paid physicians were male. Gastroenterologists comprised 44% (n = 22) of the 50 top paid physicians, dermatologists comprised 32% (n = 16), and the remaining 24% (n = 12) physicians were of other specialties. 19/22 gastroenterologists, 15/16 dermatologists, and 8/12 physicians of other specialty physicians were male. On average, the physicians had been out of training for 29 years. Fifty percent of the physicians’ h-index was < 20, 36% of physicians had an h-index between 20 and 40, 8% had an h-index between 40 and 60, and 6% had an h-index above 60. The median h-index is 19.5.

Of the 50 physicians identified, 37 (74%) had relevant publications. A total of 418 papers were determined to be relevant and included in our analysis. The majority of the authors with relevant papers were gastroenterologists (n = 19). The overall rate of full disclosure was 39.5%, partial disclosure was 30%, and no disclosure was 26.3%. In addition, 6.2% of papers did not have a required disclosure statement.

Gastroenterologists (n = 19) comprised the majority of the top 50 paid physicians with relevant publications (n = 186), followed by dermatologists (n = 10) who authored 138 papers. Dermatologists had the highest rate of full disclosure at 49.3% while gastroenterologists had the lowest rate of full disclosure (34.4%) (Fig. 1). The most common types of publications examined were observational studies (n = 132), while practice guidelines were the least common (n = 13). Articles classified as practice guidelines had the highest rates of full disclosure reporting at 69.2% and basic science papers had the lowest rate of disclosure reporting (0%) (Fig. 1). When examining disclosure rates per year, publications authored in 2019 had the lowest rate of full disclosure, while 2020 had the highest rate of full disclosure at 47.6% (Fig. 1).

Figure 1.

Disclosure completeness of relevant publications authored by the top 50 compensated physicians by A specialty, B study type, and C publication year.

The median h-index for authors was 19.5. To determine association with h-index, the authors categorized as h-index > 19.5 and h-index < 19.5. There was no significant difference between the accuracy of self-reported disclosures and h-index (57.7 vs 60.2; 95% CI 0.49–2.51; p-value 0.801).

Lastly, specific disclosure policies for the 145 different journals were examined. Ninety-one (62.8%) adhered to ICMJE policies, while 54 (37.2%) did not adhere to ICMJE policies. Of the 54 journals which did not adhere to ICMJE policies, conflict-of-interest disclosure requirements were variable, as described in Table 2; however, the majority of these journals did not specify a time frame for disclosure requirements. Analysis of disclosure rates broken down by conflict-of-interest policies demonstrated that papers published in journals with stated adherence to ICMJE policies had the highest rates of full disclosure at 45% (Table 3), while papers published in journals without COI policies or with poorly defined COI policies had far lower full disclosure rates at 33%.

Table 2.

Description of Journal Policies and Their Adherence to ICMJE Conflicts of Interest Disclosure Policies

	N Journal	Percent of total
Adhere to ICMJE	91	62.8%
Do not adhere to ICMJE	54	37.2%
Of the journals not adhering to ICMJE guidelines
At least 36 months prior to submission	5	3.4%
24 months prior to submission	5	3.4%
12 months prior to submission	1	0.7%
Time of initiation of project	1	0.7%
No time frame stated	34	23.4%
No time frame, $10,000 threshold	3	2.1%
COI not defined; no policies described	5	3.5%

Table 3.

Rates of Disclosure of Conflict of Interest in Relation to Adherence to ICMJE Policies

	N papers	Percent of total
Compliance status
ICMJE compliant	236	56.46
COI defined but no time frame stated	124	29.67
12 months, 24 months, 36+ months, or from beginning of project	41	9.81
No COI policies or policies not defined	12	2.87
No time frame stated, $10,000 threshold	5	1.20
Disclosure status
ICMJE compliant
Full disclosure	106	44.92
Partial disclosure	64	27.12
No disclosure	63	26.69
No statement included	3	1.27
COI defined but no time frame stated
Full disclosure	46	37.10
Partial disclosure	33	26.61
No disclosure	29	23.39
No statement included	16	12.90
12 months, 24 months, 36+ months, or from beginning of project
Full disclosure	9	21.95
Partial disclosure	17	41.46
No disclosure	11	26.83
No statement included	4	9.76
No COI policies or policies not clearly defined
Full disclosure	4	33.33
Partial disclosure	3	25.00
No disclosure	2	16.67
No statement included	3	25.00
No time frame stated, $10,000 threshold
No disclosure	5	100.00

DISCUSSION

This is the first study to examine discrepancies in rates of disclosure for physicians who received payments for multiple-indication monoclonal antibody drugs. The top 50 physicians compensated by medical industry for activities related to these drugs received a total of $18,484,228 from 2016 to 2020, and 37 of these physicians authored 418 relevant articles during this period. We found substantial inconsistencies in the rates of self-reported disclosures and financial relationships documented in the Open Payments Database, consistent with rates reported in previous studies, including a previous study by our group reporting a 37% disclosure rate for surgeons.^{11, 12, 15} No physician in our cohort disclosed all payments. Publications authored by dermatologists had the highest rate of self-disclosure, while those authored by gastroenterologists had the lowest rate. Female physicians comprised only 16% of the top 50 compensated physicians. Editorials, which include opinion and commentary pieces, had low rates of required disclosure statements. Publications with a high likelihood to impact patient care, clinical practice guidelines and interventional clinical trials, had the highest rates of self-disclosure of conflicts of interest.

We found that during the study period, the low rates of full disclosure remained largely constant. A potential explanation for the low rate of disclosures may be inconsistencies in reporting requirements between different journals, which was confirmed with our study. For example, ICMJE journals require 3 years of reporting for financial disclosures,⁹ and significant variability exists between different publishing organizations’ policies with regard to time period of reporting, means of compensation (financial vs non-financial), and requirements of editors and reviewers.^{16, 17} While we did find that journals adhering to ICMJE policies had the highest rate of disclosure, it was still disappointingly low at 45%. Furthermore, additional variability exists in recommendations for how conflicts of interest are managed to reduce bias.^{16, 17} We propose that unification of reporting requirements may increase rates of self-disclosure. Specifically, the authors endorse disclosure of all financial and non-financial relationships, review of these disclosures by the journal and investigation if a conflict of interest is thought to exist, and review of the Open Payments Database for accuracy as this is publicly available data.

While differences in conflicts of interest disclosure rates by study type have been previously noted at a single journal,¹⁰ our study is the first publication that analyzed disclosure rates across study types in multiple journals serving multiple specialties. Low rates of self-disclosure in editorials present a unique problem for judgment of bias and conflict of interest, as these articles are often opinion pieces that do not undergo the same peer-review scrutiny as research articles, and such opinions may not be supported by available data. Indeed, some journals do not consider pieces written by authors with any conflicts of interest for publication.¹⁸ We also found remarkably low rates of self-disclosure from authors of preclinical and basic studies. Definitions of what constitutes a conflict of interest may be less clear for physicians engaging in basic research than for clinical research or practice guidelines, as basic science studies rarely involve patients and identifying sources of bias from payments from medical industry may be difficult.¹⁹ However, basic science research is the foundation of translational research, and industry funding for preclinical studies has the potential to affect the direction of clinical studies and practice. This is unscored by literature showing industry-funded research producing proindustry results. While industry funding for basic and preclinical studies is an important funding source for translational discovery, disclosure of conflicts of interest in resulting articles is critical to ensure trust in the scientific process. Requiring disclosure of all financial and non-financial relationships would assist in eliminating the subjectivity inherent in determining what is deemed a relevant disclosure for basic science authors.

The findings in our study are in agreement with previous literature suggesting that the issue of inaccurate disclosure of conflicts of interests is not unique to the physicians in our cohort.^{12, 15} Additionally, while our study utilizes data on payments to US-based physicians, these concerns are not confined to the USA. Following the passage of the Sunshine Act in the USA, EU Member States adopted similar regulations on reporting of medical industry payments to EU physicians. However, a recent study in BMJ found that compliance with these regulations is variable across EU Member States with low accessibility, quality, and transparency of data compared to the US Open Payments Database²⁰. Further, a study conducted in Japanese medical journals found vast inconsistencies in COI disclosure requirements, education, and policy implementation between Japanese Association of Medical Sciences journals.²¹ These studies indicate that COI transparency and regulation consistency are widespread beyond the USA.

Industry payments to physicians are associated with higher rates of prescriptions and positive results in studies for promoted drugs;²² thus, it is critical that conflicts of interest are disclosed when publishing articles that influence clinical decision-making or the direction of translational science. Having a conflict of interest does not inherently imply that decision-making or work is biased, but rather that it is a risk factor for bias.¹⁵ While collaboration between medical industry and physicians advances medical knowledge, receipt of payment from medical industry is a potential source of bias and disclosing financial relationships in medical research allows these risk factors for bias to be recognized.

Failure to disclose conflicts of interests has far-reaching effects. Allowing recipients, may they be patients, reviewers, or the academic audience, to evaluate clinical recommendations and conclusions in the context of the financial relationship or funding source is critical. Not doing so erodes trust in the practice of physicians and scientists. Further, failure to disclose can harm the reputations or individuals. Additionally, the penalty for authors failing to adhere to disclosure guidelines can vary. The ICMJE follows COPE guidelines for investigation of potential conflicts of interest in post-publications articles. After investigation of the potential conflict of interest, the publication may under post-publication modification of the article or retraction, as deemed appropriate.

Limitations to our study include limitations inherent to the Open Payments Database. The Open Payments Database is known to have some inaccuracies, particularly in regard to classification of physician specialty;²³ however, these inaccuracies are more often reflected in the physicians entered into the database rather than the payment data.²⁴ In addition, there is some subjectivity inherent in deciding what is relevant when determining whether a conflict of interest exists, which may be reflected in the rates of self-reported disclosures themselves. Finally, as all reporting to the Open Payments Database is financial in nature, we were unable to assess self-disclosure accuracy of non-financial relationships.

CONCLUSION

Substantial inconsistencies were found between self-reported conflict-of-interest disclosures and payments which are documented in the Open Payments Database for physicians who received compensation for monoclonal antibody drugs. These findings could be related to inaccuracies in the Open Payments Database and/or due to the non-uniform disclosure policies among various journals or inability of authors to fully disclose due to options provided during the submission process. Since conflicts of interest are known to introduce bias, we therefore propose a uniform policy consisting of full financial and non-financial disclosure for all publications.

Declarations

Conflict of Interest

Dr. Alessio Pigazzi has received compensation and/or consulting fees from Intuitive, Ethicon, Covidien, Medtronic, Colospan, and Vioptix, honoraria from Xodus, and education and food/beverage fees under the sum of $200 from Coloplast and AcelRx. Dr. Jafari has received compensation and/or consulting fees from Intuitive, Covidien, Erbe, Merz, and AcelRx and is a course director for Medtronic. Dr. Jafari has received consulting fees from Intuitive, Covidien, Erbe, Merz, AcelRx, and Karl Storz and is a course director for Medtronic. Dr. Herre became an employee of Regeneron Pharmaceuticals during completion of the manuscript. The remainder of the authors report no proprietary or commercial interest in any product mentioned or concept discussed in this article.