Evaluation of Conflicts of Interest among Participants of the Japanese Nephrology Clinical Practice Guideline

Visual Abstract

Abstract

Background and objectives

Rigorous and transparent management strategies for conflicts of interest and clinical practice guidelines with the best available evidence are necessary for the development of nephrology guidelines. However, there was no study assessing financial and nonfinancial conflicts of interest, quality of evidence underlying the Japanese guidelines for CKD, and conflict of interest policies for guideline development.

Design, setting, participants, & measurements

This cross-sectional study examined financial and nonfinancial conflicts of interest among all 142 authors of CKD guidelines issued by the Japanese Society of Nephrology using a personal payment database from all 92 major Japanese pharmaceutical companies between 2016 and 2019 and self-citations by guideline authors. Also, the quality of evidence and strength of recommendations underlying the guidelines and conflicts of interest policies of Japanese, US, and European nephrology societies were evaluated.

Results

Among 142 authors, 125 authors (88%) received $6,742,889 in personal payments from 56 pharmaceutical companies between 2016 and 2019. Four-year combined median payment per author was $8258 (interquartile range, $2230‒$51,617). The amounts of payments and proportion of guideline authors with payments remained stable during and after guideline development. The chairperson, vice chairperson, and group leaders received higher personal payments than other guideline authors. Of 861 references in the guidelines, 69 (8%) references were self-cited by the guideline authors, and 76% of the recommendations were on the basis of low or very low quality of evidence. There were no fully rigorous and transparent conflicts of interest policies for nephrology guideline authors in the United States, Europe, and Japan.

Conclusions

Most of the Japanese CKD guideline recommendations were on the basis of low quality of evidence by the guideline authors tied with pharmaceutical companies, suggesting the need for better financial conflicts of interest management.

Introduction

Clinical practice guidelines are widely used to achieve patient-centered care through systematic diagnosis and treatment by health care professionals on the basis of the best available evidence (1). However, financial and nonfinancial conflicts of interest (COIs) among guideline authors and guideline-developing organizations can unduly bias the guideline content and recommendations for not only drug use (2–5) but also a definition of disease and diagnostic methods (6). The guidelines issued by prestigious professional medical societies are sometimes considered as a guidebook of the standard of care, making some doctors reluctant to deviate from the guideline recommendations (7,8), even if the recommendations were influenced by industries and may harm patients (6). Therefore, appropriate and rigorous management of COI is necessary to establish reliable clinical practice guidelines and achieve well-balanced, patient-centered care (7).

Previous studies in other clinical specialties have shown that COIs with substantial amounts of money are widespread among guideline authors in developed countries (3,9,10), including Japan (11–15). In the field of nephrology, Coyne (16) pointed out that pharmaceutical companies, particularly those marketing and manufacturing drugs for anemia and CKD, were involved in the development of the National Kidney Foundation’s guideline for anemia in patients with CKD, which potentially led to overuse of drugs due to marketing their drugs. In another case, one recent study by Chengappa et al. (17) found that almost two thirds of the Kidney Disease Improving Global Outcomes (KDIGO) guideline authors had financial COIs with pharmaceutical companies. Therefore, appropriate and rigorous strategies for the management of COIs and guideline development are especially required in nephrology.

However, there are no studies that have evaluated the COIs of clinical guideline authors or compared the policies of major nephrology guideline–developing organizations with comprehensive payment databases reported by pharmaceutical companies. This study aimed to elucidate the prevalence of COIs among Japanese CKD guideline authors and to evaluate the evidence underlying the guidelines and the COI policy of the Japanese Society of Nephrology compared with those of the United States and European countries.

Materials and Methods

Study Setting and Participants

This cross-sectional analysis examined the financial and nonfinancial COIs of Japanese CKD guideline authors using pharmaceutical company–reported payment data and authors' self-citations. We surveyed all authors of the evidence-based clinical practice guidelines for CKD 2018 issued by the Japanese Society of Nephrology in June 2018 (18). The Japanese Society of Nephrology is the most authoritative medical organization in Japan regarding the field of nephrology, with 10,597 members as of March 2018.

Data Collection and Payment Source

The guidelines and associated official institutional web pages provided data on author names, sexes, affiliations, and positions. Also, authors' COI statements, cited publications, quality of evidence, and strength and content of recommendations were extracted. To assess authors' academic performance, similar to previous studies, we used the Scopus database (https://www.scopus.com/freelookup/form/author.uri) to collect authors' h index and publication counts (12,19). We also searched authors of all publications cited in the guidelines in PubMed (https://pubmed.ncbi.nlm.nih.gov/) for English articles and Ichu-shiWeb (https://search.jamas.or.jp/) and Medical Online (https://www.medicalonline.jp/) for Japanese articles. When missing in these databases, we searched Google Scholar (https://scholar.google.co.jp/).

The Japan Pharmaceutical Manufacturers Association, the largest Japanese pharmaceutical trade association, has required all member companies to disclose payments to medical institutions and health care professionals on company websites since 2013 (20). We collected information on payments for lecturing, writing, and consulting from all 92 pharmaceutical companies affiliated with the Japan Pharmaceutical Manufacturers Association as previously described (13,21). As most societies require COI declarations for 3 years before guideline publication, it was preferable to examine the past 3 years to assess payments to authors. Additionally, we included the data for 1 year after the guideline publications as some societies require authors to update COI declarations continuously (10,22). Thus, we determined that we would collect data from 2016 to 2019.

Analyses

We performed a descriptive analysis of the demographic and payment data. The median payments by each demographic category were calculated. The Mann–Whitney U test for two categorical variables and one-way ANOVA for more than three categorical variables were used to evaluate differences between categories. To evaluate payment concentration among the guideline authors, the Gini index was calculated. The Gini index is a measure of distributional inequality, with zero for complete equality and one for complete inequality (23).

Then, a multivariable negative binomial regression analysis was conducted to examine potential factors related to the amount of money paid to guideline authors. Predictor variables included guideline authors' sex, type of affiliation, h index, and years since Japanese medical licensure.

One assessment of nonfinancial COI is that of authorship of original or review articles that directly affect the recommendations of guidelines (22). We performed descriptive analyses on the guideline authors and the articles self-cited by authors in the guidelines. The method concerning self-citation was described in our previous study (12).

Furthermore, recommendations and quality of evidence of the guidelines were evaluated. Similar to KDIGO guidelines, the guidelines issued by the Japanese Society of Nephrology provide four scales to evaluate the quality of evidence (in effect estimates, A for high confidence, B for moderate confidence, C for low confidence, and D for very low confidence) and three scales to express the strength of recommendations (level 1 for a strong recommendation, level 2 for a weak recommendation, and ungraded for no clear recommendation) (24).

Finally, to clarify the characteristics of COI management strategies in Japan compared with other developed countries, we collected COI policies for guideline development from the Japanese Society of Nephrology (25), the American Society of Nephrology (26,27), the National Kidney Foundation (28), the European Renal Association–European Dialysis and Transplant Association (29), the KDIGO, the US National Academy of Medicine (1), and the Guidelines International Network (30). Then, COI policies of each society were reviewed. The assessment items for COI policies were selected on the basis of previous studies and current global standard policies presented by the US National Academy of Medicine and the Guidelines International Network (1,2,22,30–33). When necessary, we contacted the societies’ guideline development committee and professional offices responsible for publishing these guidelines to confirm details of the development process, COI statements, and policies.

Japanese yen (¥) was converted into US dollars using 2016, 2017, 2018, and 2019 average monthly exchange rates of ¥108.8, ¥112.1, ¥110.4, and ¥109.0 per $1, respectively. All analyses were conducted using Microsoft Excel, version 16.0 (Microsoft Corp.) and Stata, version 15 (StataCorp).

Ethical Clearance

The ethics committee of the Medical Governance Research Institute approved this study. Informed consent was waived, and direct contact to the societies was allowed by the ethics committee of the Medical Governance Research Institute.

Results

We identified 142 guideline authors. Their sociodemographic characteristics are shown in Table 1. Table 2 summarizes the financial characteristics of personal payments from pharmaceutical companies to guideline authors. In total, 125 guideline authors (88%) received personal payments of $6,742,889 from 56 pharmaceutical companies between 2016 and 2019. The payment amounts and percentage of guideline authors receiving payments were stable between 2016 and 2019. Four-year combined average and median payments per author were $47,485 (SD: $82,229) and $8258 (interquartile range [IQR], $2230‒$51,617), respectively. For payments during 2016–2017 guidelines development, 87% (123 of 142) received a median of $3444 ($780‒$23,359) in 2-year combined payments. Similarly, for 2018–2019 payments after guideline publication, 82% (117 of 142) received a median of $4381 (IQR, $1016‒$29,225) of personal payments. The guideline chairperson and vice chairperson received personal payments of $231,024 and $233,424 over the 4-year period, which were the 10th and ninth largest payments among all 142 authors, respectively. The Gini index was 0.74 for total payments over the 4-year period, suggesting that large payments were concentrated among certain guideline authors.

Table 1.

Demographic characteristics of Japanese CKD clinical practice guideline authors

Variables	Values
Sex, n (%)
Men	122 (86)
Women	20 (14)
Affiliation, n (%)
University hospital	121 (85)
Professor	28 (20)
Nonprofessor	93 (66)
General hospital	17 (12)
Clinic	4 (3)
Medical license, n (%)
Yes	141 (99)
No	1 (1)
Clinical and academic experience, IQR
Median years after medical licensure	23 (18‒28)
Median h index	12 (7‒24)
Median no. of academic articles	43 (21‒92)
Specialty, n (%)
Nephrology	134 (94)
Public health and clinical epidemiology	5 (4)
Pediatrics	2 (1)
Urology	1 (1)
Type of involvement, n (%)
Chair and deputy chair	2 (1)
Working group leaders and deputy leaders	30 (21)
Advisor	1 (1)
Committee members	109 (77)

IQR, interquartile range.

Table 2.

Financial conflicts of interest between the Japanese CKD clinical practice guideline authors and pharmaceutical companies

Variables	2016	2017	2018	2019	Combined Total
Authors reported conflicts of interest, n (%)	Not reported	Not reported	Not reported	Not reported	Not reported
Total payments, US$	1,357,744	1,648,117	1,853,458	1,883,570	6,742,889
Average payments (SD), US$	9562 (17,877)	11,606 (21,366)	13,053 (23,061)	13,265 (23,024)	47,485 (82,229)
Median payments (IQR), US$	1765 (352‒9369)	1924 (456‒12,538)	2194 (353‒13,578)	2070 (508‒14,827)	8258 (2230‒51,617)
Payment range, US$	0‒90,545	0‒133,588	0‒120,348	0‒101,686	0‒413,332
Authors with payments, n (%)	112 (79)	113 (80)	109 (77)	111 (78)	125 (88)
Authors with payments >$10,000, n (%)	35 (25)	43 (30)	41 (29)	42 (30)	65 (46)
Authors with payments >$50,000, n (%)	8 (6)	11 (8)	15 (11)	12 (9)	37 (26)
Authors with payments >$100,000, n (%)	0 (0)	1 (1)	2 (1)	1 (1)	23 (16)
Authors with payments >$200,000, n (%)	0 (0)	0 (0)	0 (0)	0 (0)	11 (8)
Gini index	0.76	0.76	0.75	0.75	0.74

IQR, interquartile range.

There were two authors with a missing value in the year since receiving a medical license and one author with a missing value in the h index. Among guideline authors, women had lower median personal payments than men ($3383 versus $9658, respectively; P=0.009 in the Mann–Whitney U test) and lower relative monetary value (RMV) of 0.27 (95% confidence interval [95% CI], 0.12 to 0.63) (Supplemental Table 1). The working group leaders ($92,044) and guideline chairs ($232,224) had larger median payments than those with no leadership role ($5599; P<0.001). Also, the h index of guideline authors (RMV: 1.0; 95% CI, 1.0 to 1.1) and type of engagement in guideline development, such as working group leaders (RMV: 3.0; 95% CI, 1.7 to 5.5) and guideline chairs (RMV: 3.7; 95% CI, 2.0 to 7.0), had positive associations with 4-year combined total personal payments.

We identified 861 references, 808 (94%) English and 53 (6%) Japanese articles, in the Japanese Society of Nephrology CKD guidelines. For 11 Japanese guidelines cited in the CKD guidelines, author names could not be found because author disclosure was limited to subscribers. Among 142 guideline authors, 83 (58%) were authors of cited articles in the guidelines. Of 861 references, 69 (8%) references were self-cited by guideline authors (Table 3). The median number of self-cited articles per guideline author was 1.0 (IQR, 0–2.8), and the maximum was 17. Of the self-cited articles, 38% were observational studies, 25% were clinical trials, and 7% were review articles. There was no disclosure of nonfinancial COIs, such as the authorship of original and review articles in the guidelines.

Table 3.

Self-citations by the Japanese CKD clinical practice guideline authors in the guidelines

Variables	Number of References
Total references, n	861
Language of cited articles, n (%)
English	808 (94)
Japanese	53 (6)
No. of author self-cited articles in the guideline, n (%)	83 (59)
No. of self-cited articles, n (%)
Total	69 (8)
Median per author	1 (0)
Interquartile range per author	0‒2.8 (0‒0)
Range per author	0‒17 (0‒2)
Type of self-cited articles, n (%)
Observational studies (prospective and retrospective)	26 (38)
Guideline	18 (26)
Clinical trial	17 (25)
Review article	5 (7)
Systematic review and meta-analysis	2 (3)
Modeling study	1 (1)
Case report	1 (1)

The guideline had 80 clinical questions and 106 recommendations (Table 4). Of 106 recommendations, 84 (79%) were treatment recommendations, 13 (12%) were diagnostic recommendations, and nine (9%) were disease recommendations. Thirty-four (32%) were strong, 56 (53%) were weak, and 16 (15%) were ungraded recommendations. For the quality of evidence, five (5%) were high, 19 (18%) were moderate, 48 (45%) were low, and 33 (31%) were very low. Thus, 76% of recommendations were on the basis of the low or very low quality of evidence. Furthermore, only 12% of the 34 strong recommendations were on the basis of high-quality evidence, whereas a total of 16 (47%) strong recommendations were on the basis of low (35%) or very low (12%) quality of evidence. There were only four strong recommendations with high quality of evidence, including three treatment recommendations and one diagnostic recommendation (Supplemental Table 2).

Table 4.

Quality of evidence by the strength of recommendation

Quality of Evidence	Strength of Recommendations, n (%)
	Strong Recommendations (Level 1)	Weak Recommendations or Suggestions (Level 2)	Ungraded Recommendation	Overall
High (A)a	4 (12)	1 (2)	0 (0)	5 (5)
Moderate (B)b	13 (38)	6 (11)	0 (0)	19 (18)
Low (C)c	12 (35)	33 (59)	3 (19)	48 (45)
Very low (D)d	4 (12)	16 (29)	13 (81)	33 (31)
Ungraded	1 (3)	0 (0)	0 (0)	1 (1)
Total	34 (32)	56 (53)	16 (15)	106

^aHigh confidence in effect estimates.

^bModerate confidence in effect estimates.

^cLow confidence in effect estimates.

^dVery low confidence in effect estimates.

For the COI policy review, we found that no COI policies of Japanese, US, and European nephrology societies fully covered sufficient criteria for guideline development (Supplemental Table 3). Although KDIGO has published many guidelines for nephrologic diseases, this organization did not have its own COI policy and managed COIs of guideline authors following US National Academy of Medicine standards in several guidelines (34,35). The strictness of COI policies differed by the organizations. The European Dialysis and Transplant Association did not limit the proportion of guideline authors with COIs but highlighted more transparent approaches to guideline development, indicating that all past, present, and future financial COIs should be disclosed by monetary category and year received. The National Kidney Foundation published KDIGO guidelines under less strict COI policies compared with the Guidelines International Network and the US National Academy of Medicine. Meanwhile, the Japanese Society of Nephrology followed less transparent management strategies in the guideline development process, such as waiving to declare financial COIs less than a certain monetary threshold, no restriction on guideline authors’ financial activities with pharmaceutical companies during guideline development, and no patient involvement.

Furthermore, the Japanese CKD guidelines did not disclose financial COIs of guideline authors and funding sources of the Japanese Society of Nephrology, although the previous Japanese Society of Nephrology’s COI policies issued in 2017 required it to do so. In responses from the Japanese Society of Nephrology to us, the society stated that the guideline published in 2018 was out of the scope of the COI policy because 2018 was a transition period for the COI policy. No authors with COI were excluded from the deliberations for guideline recommendations. Therefore, we could not assess the accuracy of the self-reported financial COIs of the guideline authors.

Discussion

This study revealed that 88% of Japanese CKD guideline authors received a median personal payment of $8258 (IQR, $2230‒$51,617) from pharmaceutical companies between 2016 and 2019. Leading guideline authors received larger payments than others. Further, 8% of references were self-cited by guideline authors, and 76% of recommendations were on the basis of low or very low quality of evidence. The financial and nonfinancial COIs were regulated by less stringent policies in Japan than the global standard COI policies.

Substantial monetary values and a high prevalence of financial COIs among the Japanese CKD guideline authors were similar to our previous studies among Japanese guideline authors in other fields (11–14,36–38). The prevalence of Japanese guideline authors with payments ranged from 78% for oncology guidelines (37) to 95% for hematology guidelines (14), and the average annual payments ranged from $10,565 for the oncology guidelines (37) to $28,371 for infectious disease guidelines (38). However, Japanese authors of CKD guidelines may have had more financial COIs compared with those in other countries, as 66% of KDIGO guideline authors self-reported financial COIs with pharmaceutical companies (17).

Because there was no difference in the prevalence of guideline authors with financial COIs or the median amount of personal payments during guideline development, the year of publication, or 1 year after publication, we assume that the Japanese CKD guideline authors made little effort to avoid financial COIs related to guideline development. There would be several reasons for this. First, the Japanese Society of Nephrology's policy only requires the guideline authors to declare personal payments of $4587 (¥500,000) or above per year for lecturing and writing per company. In this study, the percentage of authors with $4587 (¥500,000) or more ranged from 32% in 2016 to 41% in 2018. Given these results, many guideline authors with annual personal payments of <$4587 (¥500,000) may have considered that their financial relationships with pharmaceutical companies would not be an issue and did not need to be eliminated. Meanwhile, among the top 10% of guideline authors in total payments, the median personal payment ranged from $54,121 in 2016 to $81,316 in 2019. This indicates that the guideline authors with significant financial COIs also had not made efforts to reduce their financial relationships during the guideline development. This hypothesis would be backed up by the previous studies assessing financial COIs among Japanese guideline authors of other specialties (11,12,15), as the studies also revealed that the prevalence and monetary values were constant during guideline development. Considering that several authoritative international societies, such as the European Dialysis and Transplant Association, require guideline authors to minimize their financial COI during guideline development, Japanese guideline authors and guideline-developing organizations should revise their COI policies to focus on financial COIs during and after guideline development.

For nonfinancial COIs, only 8% of references were self-cited by the Japanese CKD guideline authors, although there was no disclosure or regulation of nonfinancial COIs in the guideline. A study that assessed self-citation in the most recent Japanese guidelines for allergic rhinitis reported that guideline authors self-cited 28% of references (12), more than half of which were written in Japanese. Compared with this study, our findings suggest that the Japanese CKD guidelines were developed through a more rigorous and systematic review of evidence. However, we found that few nephrology societies have focused on nonfinancial COIs in their policies.

Furthermore, 76% of the guideline recommendations were made upon low or very low quality of evidence, and 47% of strong recommendations were supported by low or very low quality of evidence, such as a few cross-sectional studies, expert opinions, and no evidence. In previous studies in the United States, Alseiari et al. (24) reported that 55% of KDIGO recommendations were supported by low and very low quality of evidence. Further, they found that 37% of the strong recommendations were on the basis of low or very low evidence (24). Recent studies assessing the quality of evidence underlying clinical practice guidelines elucidated that many specialties (39–45) suffered from developing guidelines on the basis of low quality of evidence. One recent empirical study showed that expert consensus–based recommendations tended to be strongly stated and more inappropriate than evidence-based recommendations, even though the recommendations were on the basis of low quality of evidence (41). Special caution for the development process should be paid in developing recommendations with little high-quality evidence and many expert opinions, as different guideline authors tended to make different and sometimes quite opposite recommendations depending on lower quality of evidence, as in the cases of routine mammography for breast cancer (7,46,47) and direct-acting antiviral medications for hepatitis C (31,48–50). Changing the stringency of COI management depending on the quality of the evidence for the recommendations might be appropriate.

Our evaluation analysis of COI policies revealed that current approaches to increase transparency and minimize the financial influence on guideline development were limited in the field of nephrology in the United States and Japan, whereas the European Dialysis and Transplant Association has developed a more rigorous and transparent COI policy and was implementing it in guideline development. However, none of the societies' COI policies satisfied all of the US National Academy of Medicine and Guidelines International Network standards for COI management in guideline development. Deviations from global, rigorous, and transparent COI policies are not limited to nephrology guidelines. Jefferson and Pearson (31) found that US cholesterol management guidelines and US hepatitis C guidelines deviated from the US National Academy of Medicine standards, and this deviation remained (49). Also, Chengappa et al. (17) reported that 90% of KDIGO guidelines had more than half of the guideline authors with reported COIs, which is one of the deviations from the US National Academy of Medicine. All medical societies and organizations developing guidelines should pay more attention to financial COIs with pharmaceutical companies during guideline development and broader nonfinancial COIs. Identifying the lack of adherence to the global guideline development process and the constant review and revision of COI policies, particularly those regarding transparency and strictness in COI among the guideline authors, is essential in the field of nephrology and beyond (51).

This study includes several limitations. First, we did not analyze the inference of causality of financial COIs, nonfinancial COIs, level of evidence, and COI policy of each society. Second, we manually searched and reviewed the COI policy of each guideline-developing organization. Despite our careful review and direct inquiries to organizations, we could not exclude errors, such as omissions, in our review of COI policies. Third, our payment database was manually collected from 92 pharmaceutical companies, which may cause some errors in the process. Fourth, the payments for research were not disclosed on the individual guideline author level by the pharmaceutical companies in Japan. Therefore, this study underestimates the magnitude and prevalence of all financial relationships with pharmaceutical companies among the Japanese CKD guideline authors. Lastly, this study assessed all authors of a single Japanese CKD guideline, and the findings may not apply to other countries.

In conclusion, this study found that about 90% of Japanese CKD guideline authors had financial COIs with pharmaceutical companies and that >75% of recommendations were on the basis of low or very low quality of evidence. Furthermore, the COI policies of the Japanese Society of Nephrology and those of other nephrology societies in the United States and Europe deviated from global, rigorous, and transparent COI policies, suggesting that COI management should be improved not only in Japan but also in other countries.

Disclosures

A. Ozaki and T. Tanimoto received personal fees from Medical Network Systems outside the scope of the submitted work. H. Saito received personal fees from Taiho Pharmaceutical Co. Ltd. outside the scope of the submitted work. T. Tanimoto received personal fees from Bionics Co. Ltd. outside the scope of the submitted work. All remaining authors have nothing to disclose. Regarding intellectual COIs among the study authors, all are engaged in ongoing research examining financial and nonfinancial COIs among health care professionals and pharmaceutical companies in Japan and the United States. Individually, the authors have contributed to several published studies addressing COIs and quality of evidence among clinical practice guideline authors in Japan and the United States.

Funding

This study was funded in part by the Medical Governance Research Institute. This nonprofit enterprise receives donations from a dispensing pharmacy, namely Ain Pharmacies, Inc.; other organizations; and private individuals. This study also received support from Tansa (formerly known as the Waseda Chronicle), an independent nonprofit news organization dedicated to investigative journalism. There was no grant number for the donation to the Medical Governance Research Institute.

Author Contributions

Y. Kaneda, A. Murayama, A. Ozaki, T. Tanimoto, K. Yamada, and M. Yoshida conceptualized the study; A. Murayama, K. Yamada, and M. Yoshida were responsible for data curation; A. Murayama, A. Ozaki, and K. Yamada were responsible for formal analysis; A. Ozaki and T. Tanimoto provided supervision; Y. Kaneda, A. Murayama, A. Ozaki, H. Saito, T. Sawano, R. Shrestha, S. Shrestha, T. Tanimoto, and M. Yoshida wrote the original draft; and Y. Kaneda, A. Murayama, A. Ozaki, H. Saito, T. Sawano, R. Shrestha, S. Shrestha, T. Tanimoto, K. Yamada, and M. Yoshida reviewed and edited the manuscript.

Data Sharing Statement

All data are included in the manuscript and/or supporting materials.

Supplemental Material

This article contains the following supplemental material online at http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN.14661121/-/DCSupplemental.

Supplemental Table 1. Median monetary values and multivariable negative binomial regression models between the guideline authors’ demographic characteristics and the monetary values of the personal payments.

Supplemental Table 2. Level of evidence and strength of recommendation by content category.

Supplemental Table 3. Conflict of interest policy review between major guideline-developing organizations in Japan, the United States, and Europe.

Supplemental Table 4. Anonymized raw payment data reported by pharmaceutical companies between 2016 and 2019.