BMJ
Confidential: For Review Only
Collaboration between academics and industry in clinical trials: cross-sectional study of publications and survey of lead academic authors
Journal: BMJ
Manuscript ID BMJ.2018.044617
Article Type: Research
BMJ Journal: BMJ
Date Submitted by the Author: 17-Apr-2018
Complete List of Authors: Rasmussen, Kristine; Nordic Cochrane Centre, Rigshospitalet Dept. 7811 Bero, Lisa; University of Sydney, Charles Perkins Centre / Pharmacy Redberg, Rita; University of California San Francisco, UCSF Division of Cardiology Gøtzsche, Peter; Nordic Cochrane Centre, Rigshospitalet Dept. 7811 Lundh, Andreas; Center for Evidence-Based Medicine, Odense University Hospital and University of Southern Denmark
industry, conflicts of interest, funding, academic, authorship, ghost Keywords: authorship
https://mc.manuscriptcentral.com/bmj Page 1 of 42 BMJ
1 2 3 4 Collaboration between academics and industry in clinical trials: cross-sectional study of 5 publications and survey of lead academic authors 6 7 8 Word count: 3612 9 10 Kristine Rasmussen1 (researcher, MD, MSc), [email protected] 11 Confidential: For Review Only 12 Lisa Bero2 (professor, PhD), [email protected] 13 14 Rita Redberg3 (professor, MD, MSc, FACC), [email protected] 15 16 Peter C. Gøtzsche1 (professor, MD, DMSci), [email protected] 17 18 Andreas Lundh4 (senior researcher, MD, MSc, PhD), [email protected] 19 20 21 22 1 23 Nordic Cochrane Centre, Rigshospitalet, Copenhagen, Denmark
24 2 25 Charles Perkins Centre and Faculty of Pharmacy, University of Sydney, Australia 26 3 UCSF Division of Cardiology, 505 Parnassus Ave., Suite M 1180, San Francisco, California, USA 27 28 4 Center for Evidence Based Medicine, Odense University Hospital and University of Southern Denmark, 29 30 Odense, Denmark 31 32 33 34 § Contact information for corresponding author: Kristine Rasmussen Nordic Cochrane Centre, 35 36 Rigshospitalet Dept. 7811, Blegdamsvej 9, DK 2100 Copenhagen Ø, Denmark, 37 38 Phone: +45 35 45 7110, Fax: +45 35 45 7007, E mail: [email protected] 39 40 41 42 Competing interests: All authors have completed the ICMJE uniform disclosure form 43 at www.icmje.org/coi_disclosure.pdf and declare: no support from any organisation for the 44 45 submitted work; RFR is editor of JAMA Internal Medicine which is included in our sample of 46 journals, but RFR had no role in data extraction or analysis of the results; all authors declare no 47 48 financial relationships with any organisations that might have an interest in the submitted work in 49 the previous three years and no other relationships or activities that could appear to have 50 51 influenced the submitted work. 52 53 Funding: No external funding received. Study was financed by institutional salaries. 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 2 of 42
1 2 3 4 BOX: 5 6 Section 1: What is already known on this subject 7 8 The biomedical industry funds most clinical trials. Industry funders may influence how trials are 9 designed and reported, sometimes serving financial rather than public interest. 10 11 SectionConfidential: 2: What this study adds For Review Only 12 13 All aspects of the clinical trial were influenced by the industry funder in most industry funded trials 14 in high impact medical journals. Lead academic authors involved in industry funded trials found the 15 collaboration with the funder beneficial, but some reported loss of academic freedom. 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 3 of 42 BMJ
1 2 3 4 ABSTRACT 5 6 Word count: 291 7 8 Objective: To determine the role of academic authors, funders and contract research 9 organisations in industry funded vaccine, drug and device trials, and to determine lead academic 10 11 authors’Confidential: experiences with industry funder collaborations.For Review Only 12 13 Design: Cross sectional analysis of trial publications and survey of lead academic authors. 14 15 Eligibility criteria for selecting studies: We included the most recent 200 phase III and IV 16 17 vaccine, drug and device trials with full industry funding, at least one academic author, published in 18 one of the top seven high impact general medical journals (NEJM, Lancet, JAMA, BMJ, Annals of 19 20 Internal Medicine, JAMA Internal Medicine and PLoS Medicine). 21 22 Results: Employees of industry funders co authored 173 (87%) of publications. We found 183 23 24 (92%) trials reported funder involvement in design and 167 (84%) academic author involvement. 25 Data analysis involved the funder in 146 (73%) trials and the academic authors in 79 (40%). Trial 26 27 reporting involved the funder in 173 (87%) trials and academic authors in 197 (99%). Contract 28 research organisations were involved in the reporting of 123 (62%) trials. 29 30 Eighty of 200 lead academic authors (40%) responded to the survey. Twenty nine (33%) of the 80 31 32 responders reported that academics had final say on the design. Ten responders described 33 34 involvement of an unnamed funder and/or CRO employee in the data analysis and/or 35 reporting. Most academic authors found the collaboration with industry funder beneficial, but 3 36 37 (4%) experienced delay in publication due to the industry funder and 9 (11%) reported 38 disagreements with the industry funder mostly concerning trial design and reporting. 39 40 41 Conclusions: Industry employees and academic authors are involved in the design, conduct, and 42 reporting of most industry funded trials in high impact journals. However, data analysis is often 43 44 conducted without academic involvement. Academics view the collaboration as beneficial, but 45 some report loss of academic freedom. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 4 of 42
1 2 3 4 INTRODUCTION 5 6 Collaboration between industry and academics is common in vaccine, drug and device 7 8 development, as it can be mutually beneficial. The academics provide access to trial participants 9 and clinical and methodological expertise, and industry provides funding and expertise. The degree 10 11 of independenceConfidential: and roles of academics and For industry Review varies across trials. OnlyTrials may be 12 completely run by academic trial units with unrestricted industry funding or solely provision of study 13 medication.1, 2 Alternatively, academics are involved in trials as honorary authors in order to give a 14 15 trial scientific credibility and downplay the role of industry funder.3, 4 16 17 Based on previous work and examples, collaboration between academics and industry may result 18 19 in commercial considerations outweighing science and constraints on academic freedom. For 20 example, previous studies have found some trial agreements permit funders to block publication.5 8 21 22 Furthermore, academics have reported problems with stalling of publications, restriction of 23 publication rights and threats to never fund their institution again due to reporting of negative 24 9 25 results or harms. A survey of Canadian trial investigators found problems in relation to industry 26 control over study design, data ownership, data access and analysis.10 27 28 Although most clinical drug and device trials are industry funded, the nature of the collaboration 29 30 between industry and academics has received little attention.11 Previous studies have addressed 31 the issue on a general level.5 10, 12 To our knowledge, no contemporary studies have surveyed the 32 33 role and practices of academics, industry, and contract research organisations (CROs) when 34 collaborating in clinical vaccine, drug or device trials. 35 36 37 38 39 Objective 40 41 • To determine the role of academic authors, funders and CROs in industry funded vaccine, 42 43 drug and device trials. 44 • To determine lead academic authors’ experiences with industry funder collaborations. 45 46 47 48 METHODS 49 50 51 This study was in two parts: 1) a cross sectional study of trial publications and 2) a survey of the 52 lead academic authors of the 200 most recent industry funded vaccine, drug and device trials 53 published in one of the top seven high impact general medical journals. 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 5 of 42 BMJ
1 2 3 4 5 6 Cross-sectional Study 7 8 Search 9 10 11 One authorConfidential: (KR) manually searched the top For seven general Review and internal medical Only journals according 12 to the impact factor of the 2015 Journal Citation Report (New England Journal of Medicine (NEJM), 13 14 Lancet, JAMA, BMJ, Annals of Internal Medicine, JAMA Internal Medicine and PLoS Medicine) for 15 the most recent trial publications meeting our inclusion criteria.13 16 17 18 19 20 Inclusion criteria 21 22 We included publications of phase III and IV trials with one or more academic authors (determined 23 using the institutional address) that disclosed full industry funding. We defined full industry funding 24 25 as any trial with funding exclusively from one or more vaccine, pharmaceutical or device 26 companies. We included industry supported trials described as investigator initiated if the only 27 14 28 funding was from industry sources. 29 30 We defined academic authors as authors whose affiliation in the publication was a clinic, hospital, 31 32 university or non profit academic research centre. The lead academic author was selected 33 according to the following rank: corresponding author, first author, last author, second author, third 34 35 author, etc. If there were more than one academic corresponding author, the lead academic would 36 be deemed according to the following rank: first author, last author, second author, third author, 37 38 etc. 39 40 41 42 Exclusion criteria 43 44 • Trials of fluid therapy and dietary supplements. 45 46 • Trials with mixed funding (public and industry funded, including free provision of study drug 47 or device only). 48 49 • Secondary publications (e.g. subgroup analysis). 50 If two publications of the same trials were identified (e.g. an interim analysis and the 51 • 52 planned analysis) only the most recent publication was included. 53 54 • If a lead academic author had published more than one eligible trial, we only included the 55 most recent trial. 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 6 of 42
1 2 3 4 A second author (AL) confirmed that the 200 trial publications met the inclusion and exclusion 5 criteria. Disagreements were resolved by discussion. 6 7 8 9 10 Data extraction 11 Confidential: For Review Only 12 For each trial publication, two authors (AL, KR) independently extracted data into a standardised 13 14 data sheet. We extracted data on the role of the academic authors, funder and CRO with regard to 15 trial design, conduct, data analysis and reporting. We also extracted data on access to data, trial 16 17 agreements and the lead academic author’s reported conflicts of interest (COIs). Being an 18 investigator or receiving a grant from the trial funder in relation to the included trial was not 19 20 considered a COI (Appendix for full details). Disagreements were resolved by discussion. 21 22 During the initiation of the data extraction, we found that the contribution of individual authors and 23 role of the funder was often not clear from the publications in NEJM, as roles were typically 24 25 described in broad terms (e.g. the trial was designed by a steering committee of academics and 26 funder representatives, but the publication did not describe which authors were members). We 27 28 therefore decided to include supplementary material (e.g. lists of steering committee members and 29 contributions statements) for data extraction of NEJM publications, if available. 30 31
32 33 34 Survey 35 36 We designed our survey using Qualtrics®.15 Each lead academic author was tracked by a unique 37 38 identifier to allow matching of the survey response to the publication and in order to track authors 39 who had not responded. Results were reported in an anonymised form and the identity of the 40 academic authors was only known to two of us (AL, KR). 41 42 43 Survey questions were designed, ordered and sent out in a fashion recommended for increasing 44 the response rate and the reliability of the responses.16, 17 The survey included 20 questions with 45 46 possible answers in tick boxes and the possibility of writing a more detailed comment. The 47 questions addressed design, analysis and reporting of the trial, data access, trial agreements, and 48 49 experience with the collaboration (Appendix). 50 51 We pilot tested the survey on seven academics identified from our network who had previously 52 collaborated with industry and modified the survey slightly according to their advice. 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 7 of 42 BMJ
1 2 3 4 We emailed the survey to the 200 lead academic authors in June 2017 and reminders were 5 emailed twice monthly until the end of September 2017 (seven reminder emails). Thereafter, we 6 7 contacted non responders by letter in October 2017 and by phone in November 2017. 8 9 If a survey response had a conflicting comment, e.g. the lead academic author had chosen 10 11 “funder”Confidential: in response to a question while it wasFor clear inReview the comments that theOnly academic authors 12 were also involved, this would be reclassified after discussion between two authors (AL, KR). 13 Responses were reclassified due to comments for 16 responders. 14 15 16 17 18 Data analysis 19 20 We used only descriptive statistics. Responder comments were reported in a collated fashion 21 22 where possible. Discrepancies between survey and trial publication were reported. For trial 23 characteristics, we also stratified characteristics by vaccine, drug and device trials and survey 24 25 response type (i.e. responders, declined via email, accessed survey, but did not respond, and non 26 responders). 27 28 29 For the survey, we stratified responses by vaccine, drug and device trials. We also stratified survey 30 responses into those still collaborating with the funder or who would collaborate in the future and 31 those no longer collaborating with the funder to assess whether these had important differences. 32 33 34 For both trial characteristics and survey responses, we stratified results into groups of independent 35 trials and industry trials as we hypothesised that these groups would differ. Trials were classified 36 37 as independent if neither funder nor a CRO was involved in design, conduct, analysis, or reporting. 38 39 To assess the impact of COIs in relation to two survey questions on “funder problems” we stratified 40 responses by lead academic authors with COIs related to the funder and by lead academic authors 41 42 without COIs. 43 44 45 46 47 RESULTS 48 49 Cross-sectional study 50 51 We identified 1139 publications and included 200 trials published between July 2014 and April 52 53 2017 (Figure 1). We included trials from all selected journals except PLoS Medicine where no trials 54 met our inclusion criteria because none were fully industry funded. 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 8 of 42
1 2 3 4 5 6 Trial characteristics 7 8 All trials but 7 were published in NEJM (53%), Lancet (31%) and JAMA (13%) and 165 trials (83%) 9 10 were drugConfidential: trials (Table 1 and Appendix for stratifiedFor analyses).Review The median Only number of authors was 11 19 (range 5 103), with one trial, published in JAMA, having a study group listed as authors, which 12 13 consisted of 103 people. In 173 (87%) of the trials, there was one or more co authors who were 14 employees of the funder. The corresponding author was an academic author in 192 (96%) trials. 15 16 The lead academic author had COIs with the trial funder in 165 (83%) trials. 17 18 19 20 Involvement of funder, academic authors and CROs as reported in trial publications 21 22 23 In 137 (69%) trials, both funder and academic authors were involved in the design and in 129 24 (65%) trials in conduct (Figure 2). In 109 (55%) trials, data analysis was done by the funder and/or 25 26 a CRO without the involvement of academic authors and only 26 (13%) trials had data analysis 27 done solely by academic authors. In 117 (59%) trials, the reporting involved the funder, academic 28 29 authors and a CRO, primarily a medical writer. Only 8 (4%) trials were classified as independent 30 trials (i.e. all aspects of the industry funded trial were carried out by academic authors without 31 involvement of the funder or a CRO). In 4 of these 8 trials the lead academic author had disclosed 32 33 COIs with the funder. 34 35 The data analysis was done by a median of 3 (range 1 31) authors with the involvement of a 36 37 median of 1 (range 0 11) funder employee. In 95 trial publications, we were unable to identify the 38 exact number of authors involved in data analysis. Although we aimed to extract data on who 39 40 performed the actual statistical analysis, this was often difficult, as publications used phrases such 41 as “All authors analysed and interpreted the data” even though a publication might have 18 42 43 authors. Other publications used a phrase from the ICMJE requirements for authorship describing 44 individual authors involvement in “Acquisition, analysis, or interpretation of data“, and thus grouped 45 46 three very different processes making it impossible to assess who performed the statistical 47 analysis.(18) In 10 trials, it was stated that the funder had no role in the data analysis, but it was 48 49 unclear who else was involved (e.g. a CRO). 50 51 52 53 Data access and trial agreements 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 9 of 42 BMJ
1 2 3 4 All authors declared access to data in 77 (39%) trials and some academic authors did so in an 5 additional 67 (34%) trials. For the remaining trials, 52 (26%) stated that authors vouched for the 6 7 data without specifying who had access and 4 (2%) had no description of data access. The type of 8 data available to authors was not specified in 135 (94%) of the 144 trials reporting data access, 2 9 10 (1%) reportedConfidential: access to raw data, 3 (2%) accessFor to analysedReview data only, 1 (1%)Only access to summary 11 data for all patients, but only individual patient data from own institution, and 3 (2%) access to data 12 13 used in the publication. 14 15 In 16 (8%) trial publications, confidentiality agreements with the funder were described and 4 (2%) 16 publications described that there were no such agreements. For the remaining 180 (90%) trials, 17 18 there was no mention of confidentiality agreements in the publication. 19 20 21 22 Survey 23 24 25 We received responses from 106 (53%) lead academic authors (Figure 1). Three lead academic 26 authors were unreachable by email, post, and telephone. Eighty out of 200 (40%) responded to 27 28 survey questions, 10 (5%) accessed the survey without providing any response and 16 (8%) 29 replied that they did not want to participate (Appendix). Our comparison of trial and author 30 31 characteristics between those who responded to those who did not, revealed no important 32 differences (Appendix). 33 34
35 36 37 Decisions and involvement in trial design, analysis and reporting, as reported by lead academic 38 authors 39 40 Twenty six (33%) of the 80 responders reported that academic authors solely decided which 41 42 comparator treatment to include whereas this was only the case for choice of outcomes in 4 (5%) 43 of the responders’ trials (Figure 3 for all responses, Figure 4 for selected author comments, and 44 45 Appendix for additional author comments). Twenty nine (36%) responders reported the academic 46 authors had the final say in trial design. One responder reported the funder was involved in the 47 48 design, but this was not described in the publication. Ten responders (6 from North America and 4 49 from Europe) acknowledged a drug regulatory agency as having the main decision power on 50 51 choice of comparator, outcome and/or final say on the trial design. However, the corresponding 52 publications only acknowledged regulator input in 2 of the 10 trials. 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 10 of 42
1 2 3 4 Four (5%) responders reported funder involvement in data analysis, but according to the responder 5 the funder employee was not named in the trial publication. Three (4%) other responders reported, 6 7 the funder was involved in data analysis, but this was not described at all in the trial publication. 8 9 Four (5%) responders reported CRO involvement in data analysis, but the CRO employee who did 10 11 the analysisConfidential: was not named in the publication. For Two (3%) Review other responders Onlyreported CRO 12 involvement in data analysis, while the corresponding publications did not describe this 13 involvement. 14 15 16 Fifty three (66%) responders reported that the manuscript was drafted solely by academic authors. 17 However, 27 (51%) of these responders’ trials had writing assistance from a CRO according to the 18 19 corresponding publication. Four (5%) responders reported the manuscript was drafted by a funder 20 employee who was not a co author or a named contributor in the trial publication. One responder 21 22 reported an unnamed CRO employee drafted the manuscript together with academic authors and 23 another responder did not know who drafted the manuscript. One responder reported CRO 24 25 involvement in reporting, while the corresponding publication did not describe this involvement. 26 Fifty two (65%) responders reported academic authors solely had final say on the published 27 28 manuscript. However, 40 of these 52 trials included funder employed co authors. 29 30 In total, 10 responders described involvement of an unnamed funder and/or CRO employee in the 31 data analysis and/or reporting. An additional 7 responders described funder and/or CRO employee 32 33 involvement in trial design, data analysis, or reporting which was not reported in the corresponding 34 publication. 35 36 37 38 39 Trial agreements and data access 40 41 Sixty three (79%) responders reported a trial agreement was signed and 46 (73%) of these 63 42 43 included a publication agreement, 33 (52%) a presentation agreement and 39 (62%) a 44 confidentiality agreement (Table 2). Eight lead academic authors commented that the contract they 45 46 signed with the industry funder gave the funder the right to review and comment on the manuscript 47 and presentations, but did not give the funder approval rights. However, six of these responders’ 48 49 eight trials had funder employed co authors thereby indirectly giving funder approval right 50 (Appendix) and one of these authors also described disagreements with the funder over 51 52 manuscript content. Only 4 (6%) of these 63 trials with agreements, had a description of a 53 confidentiality agreement in the trial publication. Another trial publication reported there was no 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 11 of 42 BMJ
1 2 3 4 confidentiality agreement, but the survey contradicted this and described that the data was 100% 5 embargoed prior to publication suggesting the existence of a confidentiality agreement. 6 7 8 According to the survey, 63 (79%) of the lead academic authors had access to the entire dataset 9 and in 56 (89%) of these trials, the access was used. Seventy three of the trials allowed for a 10 11 comparisonConfidential: of data access between survey For responses Review and statements in trialOnly publications. In 5 12 (7%) trials, the lead academic author had data access according to the publication, but they denied 13 such access in the survey. In 3 trial publications that used the phrase: ”the authors vouch for the 14 15 data,” the lead academic author did not have data access according to the survey. 16 17 18 19 20 Experience with collaboration 21 22 Three (4%) responders had experienced publication delay and 9 (11%) had disagreements with 23 the funder, mostly concerning trial design and reporting (Table 2). Disagreements were generally 24 25 described as minor (Figure 3 and Appendix for detailed comments). 26 27 Sixty seven (84%) of lead academic authors were still collaborating with the industry funder or 28 29 would like to collaborate with the funder again. The most commonly reported benefit of 30 collaborating with the industry funder was funding (reported by 33 (41%) responders and by 18 31 (23%) responders as the only benefit). Thirteen (16%) responders found the fact that they were 32 33 contributing new research a benefit of the collaboration. Eight (10%) responders reported the trial 34 publication and subsequent publications using the same dataset as a benefit of the collaboration 35 36 (Figure 4 and Appendix). 37 38 In our subgroup analyses we did not find any differences in trial characteristics or survey 39 40 responses in relation to COIs, type of intervention, industry involvement and collaboration status 41 (Appendix). 42 43
44 45 46 DISCUSSION 47 48 We found that industry funded trials were mostly conducted in a collaborative fashion. 49 Nevertheless, the role of academic authors, funders and CROs varied greatly. Generally, both 50 51 academic authors and funders were involved in the trial design, conduct and reporting. However, 52 the actual data analysis was most often conducted by funder or CRO employees. Few industry 53 54 funded trials were completely independently conducted by academics and sometimes industry 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 12 of 42
1 2 3 4 involvement was downplayed or omitted in trial publications. The lead academic authors frequently 5 found the collaboration beneficial, particularly in relation to funding of the trial. However, some 6 7 academic authors reported disagreements with the funder, mostly concerning trial design and 8 reporting. 9 10 11 Confidential: For Review Only 12 13 Context 14 15 Lundh et al. analysed trials published in the Lancet in 2008 9 and had access to trial protocols.12 16 17 Our findings of extensive involvement of industry funders and that the role of funders and CROs 18 were sometimes downplayed in the trial publications are similar to the findings of Lundh et al. Data 19 20 is often stored and owned by the industry funder and similar to Lundh et al. we found that trial 21 publications rarely described what type of data the academic authors had access to and whether 22 they used this access.19, 20 We believe that descriptions of data access in publications may not be 23 24 accurate. In 67 of the 69 Lancet protocols, there was no information on academic authors’ access 25 to data, in striking contrast to the papers, which indicated that one or more academic authors had 26 27 access to the data in 64 trials.12 28 29 Ghost authorship, where individuals are involved in important aspects of clinical trials without the 30 10, 21 31 involvement being disclosed in the publications, has been well described. We found evidence 32 of ghost authorship in 17 of 80 trials, but this is likely an underestimate as some lead academic 33 34 authors had a small role in the conduct, analysis and reporting of the trial. Thus, “ghost authoring” 35 could have taken place without the lead academic authors being aware. Gøtzsche et al. found a 36 21 37 much higher prevalence of 75% of ghost authorship in trials, particular in relation to data analysis. 38 However, prevalence of ghost authorship may have diminished over time since we found that 87% 39 21 40 of trials had industry co authors compared to 64% in the study by Gøtzsche et al. 41 42 Some academic authors reported that having a high impact publication was a benefit from the 43 collaboration and academic authors were frequently (96%) prominently placed as corresponding 44 45 authors. Our findings are consistent with previous studies which found high prevalences of industry 46 ties among lead academic authors of industry funded trials.22, 23 These financial COIs may impact 47 48 on trial results. Industry funded trials by lead academic authors with company ties are more often 49 favourable compared to trials without authors’ with financial COIs and it has been suggested that 50 22, 24, 25 51 key opinion leaders with industry ties may be more willing to accept commercial pressure. 52 Previous studies have found that some COIs are under reported in publications, so the actual 53 26, 27 54 figures may be even higher. 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 13 of 42 BMJ
1 2 3 4 5 6 Strengths and weaknesses 7 8 To our knowledge, this is the first study that directly surveyed a large cohort of international 9 10 academicConfidential: authors involved in industry funded For trials published Review in high impact Only general medical 11 journals. A major strength of our study is the fact that we included trials from all over the world and 12 13 thus our results are applicable at an international level. Our usage of data from both trial 14 publications and survey responses allowed us more comprehensive data compared with studies 15 8, 10 16 solely based on a single information source. Furthermore, the anonymised format of the survey 17 is likely to have provided us with more truthful responses, especially with regard to sensitive topics. 18 19 Our findings may be somewhat limited by the retrospective nature of the survey. However, we only 20 included the most recent trials and therefore the authors are likely to remember the details 21 10 22 compared to studies focusing on past experience of academics in general. Another possible 23 limitation is the mixture of industry funded trials initiated by academics and company trials initiated 24 25 for regulatory purposes. It was not possible from trial publications to completely distinguish 26 between these types of trials and funder involvement is likely more pronounced in industry funded 27 28 trials conducted for regulatory purposes. Additionally, we only assessed the supplementary 29 material for trials published in NEJM. Academic authors with industry ties are more likely to be 30 31 favourable towards industry involvement in research than those without ties, and thus our survey 32 results may exhibit this bias.28 Finally, despite many reminders we only received survey responses 33 34 from 80 (40%) academic authors which was somewhat lower than previous surveys of academic 35 authors, but similar to Smyth et al.10, 19, 29 Nevertheless, the trial and author characteristics did not 36 37 differ amongst responders and non responders. 38 39 40 41 Meaning of the study 42 43 We found that there is room for more accurate reporting of authors’ contribution in industry funded 44 45 trials published in high impact general medical journals. A solution could be that guidelines such as 46 ICMJE and CONSORT require more detailed reporting, particularly related to data analysis, data 47 48 ownership and access.18, 30 Furthermore, journals could refuse publication when these elements 49 are not clearly reported. Trials from high impact journals have important impact on clinical 50 31, 32 51 decisions. Nevertheless, only a few of our included trials had independent analysis. However, it 52 is possible for academics to demand control over design, data storage, full data ownership, 53 54 analysis and reporting, thereby improving independence and greater reliability of trial results. 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 14 of 42
1 2 3 4 5 6 CONCLUSION 7 8 Industry employees and academic authors are involved in the design, conduct, and reporting of 9 10 most industry fundedConfidential: trials in high impact journals.For However, Review the data analysis Only was often 11 conducted without academic involvement. Academics view the collaboration with the industry 12 13 funder as beneficial, but some report loss of academic freedom. 14 15 16 17 ETHICS 18 19 We contacted the Regional Committee on Health Research Ethics and the Danish Data Protection 20 21 Agency and both confirmed that this survey does not require any form of approval according to 22 Danish law. 23 24 CONTRIBUTIONS 25 26 KR conceived the idea for the study and developed the protocol with contributions from all authors. 27 28 KR and AL extracted data and analysed the results. KR drafted the manuscript. All authors 29 interpreted results and were involved in the reporting of the study and approved the final version of 30 the manuscript. KR is the guarantor of the study and accepts full responsibility for the work and the 31 conduct of the study, had access to the data, and controlled the decision to publish. 32 33 LICENSE FOR PUBLICATION 34 35 “The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf 36 of all authors, a worldwide licence to the Publishers and its licensees in perpetuity, in all forms, 37 formats and media (whether known now or created in the future), to i) publish, reproduce, 38 distribute, display and store the Contribution, ii) translate the Contribution into other languages, 39 create adaptations, reprints, include within collections and create summaries, extracts and/or, 40 41 abstracts of the Contribution, iii) create any other derivative work(s) based on the Contribution, iv) 42 to exploit all subsidiary rights in the Contribution, v) the inclusion of electronic links from the 43 Contribution to third party material where-ever it may be located; and, vi) licence any third party to 44 do any or all of the above." 45 46 TRANSPARENCY STATEMENT 47 48 The manuscript's guarantor (KR) affirms that the manuscript is an honest, accurate, and 49 transparent account of the study being reported; that no important aspects of the study have been 50 omitted; and that any discrepancies from the study as originally planned have been explained. 51 52 53 DATA SHARING STATEMENT 54 55 Anonymised data is available upon request. 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 15 of 42 BMJ
1 2 3 4 5 REFERENCES 6 7 1. Goldenberg NA, Spyropoulos AC, Halperin JL, et al. Antithrombotic Trials Leadership and 8 Steering Group. Improving academic leadership and oversight in large industry sponsored 9 10 clinicalConfidential: trials: the ARO CRO model. Blood. For 2011;117:2089 92. Review Only 11 2. Fox L, Toms C, Kernaghan S, Snowdon C, Bliss JM. Conducting non commercial international 12 13 clinical trials: the ICR CTSU experience. Trials. 2017;18:440. 14 3. Fong EA, Wilhite AW. Authorship and citation manipulation in academic research. PLoS 15 16 One. 2017;12: e0187394. 17 4. Wislar JS, Flanagin A, Fontanarosa PB, Deangelis CD. Honorary and ghost authorship in high 18 19 impact biomedical journals: a cross sectional survey. BMJ. 2011;343:d6128. 20 5. Kneller R, Mongeon M, Cope J, Garner C, Ternouth P. Industry university collaborations in 21 22 Canada, Japan, the UK and USA with emphasis on publication freedom and managing the 23 intellectual property lock up problem. PLoS One. 2014;9:e90302. 24 25 6. Mello MM, Clarridge BR, Studdert DM. Academic medical centers' standards for clinical 26 trial agreements with industry. N Engl J Med. 2005;352:2202 10. 27 28 7. Kasenda B, von Elm E, You JJ, et al. Agreements between Industry and Academia 29 on Publication Rights: A Retrospective Study of Protocols and publications of 30 31 Randomized Clinical Trials. PLoS Med. 2016;13:e1002046. 32 8. Schulman KA, Seils DM, Timbie JW, et al. A national survey of provisions in clinical trial 33 34 agreements between medical schools and industry sponsors. N Engl J Med. 2002;347:1335 35 41. 36 9. Bodenheimer T. Uneasy alliance clinical investigators and the pharmaceutical industry. N 37 38 Engl J Med. 2000;342:1539 44. 39 10. Rochon PA, Sekeres M, Hoey J, et al. Investigator experiences with 40 41 financial conflicts of interest in clinical trials. Trials. 2011;12:9. 42 11. Moses H, Matheson DH, Cairns Smith S, George BP, Palisch C, Dorsey ER. The anatomy of 43 44 medical research: US and international comparisons. JAMA. 2015;313:174 89. 45 12. Lundh A, Krogsbøll LT, Gøtzsche PC. Sponsors' participation in conduct and reporting of 46 47 industry trials: a descriptive study. Trials. 2012;13:146. 48 13. 2015 Journal Citation Reports® Science Edition (Clarivate Analytics, 2017). (Accessed 49 50 14.01.17). 51 14. van Lent M, Overbeke J, Out HJ. Recommendations for a uniform assessment of publication 52 53 bias related to funding source. BMC Med Res Methodol. 2013;13:120. 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 16 of 42
1 2 3 4 15. Qualtrics software, Version June to December 2017 of Qualtrics. Copyright © 2018 Qualtrics, 5 Provo, UT, USA. http://www.qualtrics.com. 6 7 16. Del Greco L, Walop W. Questionnaire development: 1. Formulation. CMAJ. 1987;136:583 5. 8 17. Edwards PJ, Roberts I, Clarke MJ, et al. Methods to increase response to postal and 9 10 electronicConfidential: questionnaires. Cochrane Database For Syst Review Rev. 2009; 3:MR000008. Only 11 18. ICMJE. Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly 12 13 Work in Medical Journals Updated December 2016. http://www.icmje.org/icmje 14 recommendations.pdf (Accessed 15.12.17). 15 16 19. Lundh A, Krogsbøll LT, Gøtzsche PC. Access to data in industry sponsored trials. Lancet. 17 2011;378:1995 6. 18 19 20. Gøtzsche PC, Hróbjartsson A, Johansen HK, Haahr MT, Altman DG, Chan AW. Constraints 20 on Publication Rights in Industry Initiated Clinical Trials. JAMA. 2006;295:1641–6. 21 22 21. Gøtzsche PC, Hróbjartsson A, Johansen HK, Haahr MT, Altman DG, Chan AW. Ghost 23 authorship in industry initiated randomised trials. PLoS Med. 2007;4:e19. 24 25 22. Ahn R, Woodbridge A, Abraham A, et al. Financial ties of principal investigators and 26 randomized controlled trial outcomes: cross sectional study. BMJ 2017;356:i6770. 27 28 23. Rose SL, Krzyzanowska MK, Joffe S. Relationships between authorship contributions and 29 authors' industry financial ties among oncology clinical trials. J Clin Oncol. 2010;28:1316 21. 30 31 24. Lundh A, Lexchin J, Mintzes B, Schroll JB, Bero L. Industry sponsorship and research 32 outcome. Cochrane Database Syst Rev. 2017;2:MR000033. 33 34 25. Lundh A, Bero L. The ties that bind. BMJ. 2017;356:j176. 35 26. Cook D, Kaminski K. Competing Interest Disclosures Compared With Industry Payments 36 Reporting Among Highly Cited Authors in Clinical Medicine. International Congress on Peer 37 38 Review and Scientific Publication 2017; 12 September 2017; Chicago, USA: American Medical 39 Association; 2017. 40 41 27. Rasmussen K, Schroll J, Gotzsche PC, Lundh A. Under reporting of conflicts of interest among 42 trialists: a cross sectional study. J R Soc Med. 2015;108:101 7. 43 44 28. Glaser BE, Bero LA. Attitudes of academic and clinical researchers toward financial ties in 45 research: a systematic review. Sci Eng Ethics. 2005;11:553 73. 46 47 29. Smyth RM, Kirkham JJ, Jacoby A, Altman DG, Gamble C, Williamson PR. Frequency and 48 reasons for outcome reporting bias in clinical trials: interviews with trialists. BMJ. 49 50 2011;342:c7153. 51 30. Schulz KF, Altman DG, Moher D, for the CONSORT Group. CONSORT 2010 Statement: 52 53 updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c332. 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 17 of 42 BMJ
1 2 3 4 31. Lugtenberg M, Burgers JS, Westert GP. Effects of evidence based clinical practice guidelines 5 on quality of care: a systematic review. Quality and Safety in Health Care. 2009;18:385 92. 6 7 32. Cook DJ, Guyatt GH, Laupacis A, Sackett DL. Rules of evidence and clinical 8 recommendations on the use of antithrombotic agents. Chest. 1992;102:305S 11S. 9 10 11 Confidential: For Review Only 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 18 of 42
1 2 3 4 TABLES AND FIGURES 5 6 FIGURE 1. STUDY FLOW DIAGRAM 7 8 9 10 11 Confidential: For Review Only 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Figure 1: *A response was included by one of the authors who later replied that he did not have time to 33 34 complete the survey. 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 19 of 42 BMJ
1 2 3 4 FIGURE 2. REPORTED INVOLVEMENT IN THE 200 TRIALS 5 6 7 8 9 10 11 Confidential: For Review Only 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 CRO: contract research organisation. 35 36 37
38 39 40 41 42 43 44 45 46 47 48
49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 20 of 42
1 2 3 4 FIGURE 3. REPORTED INVOLVEMENT ACCORDING TO SURVEY 5 6 7 8 9 10 11 Confidential: For Review Only 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 CRO:contract research organisation 32 *Funder and CRO also includes unacknowledged persons for the conducted statistical analysis and drafting 33 of manuscript. 34 **Other includes regulator, journal, unknown or unclear. 35 ***Collaboration includes academic, funder, and/or CRO; academic, funder and/or regulator; academic and 36 CRO; funder and CRO. 37 No responders reported CRO as the only actor for any of the categories above. 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 21 of 42 BMJ
1 2 3 4 FIGURE 4. SELECTED COMMENTS TO THE SURVEY QUESTIONS 5 6 Selected comments from Benefits of collaborating with the trial industry funder 7 Funding: “Study would not have been possible without funding which came as a donation. No strings 8 attached.” Publication: “The Journal X paper! And more to come.” 9 “Opportunity to author papers of important study results in high impact journals” 10 PersonalConfidential: benefits: “Being PI of a positive trial is For always a benefitReview in CV and recognition”; Only 11 “ That being said, I would not object to all industry sponsored studies being listed as 'the Trial X 12 Investigators' or something similar ultimately, despite our input, these are studies done by the company” 13 Selected comments for confidentiality agreement 14 Confidential until publication: “No for results which are already in the public domain, obviously, but yes for all 15 other unpublished data” 16 Miscellaneous: “The agreement stated that sponsor could request up to 30 days confidentiality if necessary 17 for patent protection.” 18 Selected comments for choice of comparator 19 Collaboration: “The company has to obviously agree with the study design. They fund the trial.” 20 Miscellaneous: “The sponsor did not want a placebo arm in the trial. We insisted upon it for safety evaluation and they ultimately agreed to a placebo arm with deferred treatment.” 21 Selected comments for final say in study design 22 Funder: “We advised the sponsor study team (extensively) and I would generally say that they listened, but 23 they did not 'require' sign off from the steering committee before finalizing the protocol so in that sense, they 24 had final say.” 25 Selected comments for statistical analysis 26 Miscellaneous: “We analyse with independent academic and with the funder statisticians. Sometimes we 27 have to rely on only the funder statisticians which is less satisfactory. However SAPs and protocols are 28 submitted to regulatory and editors before analysis and publication.”; 29 “Stats team employed by company. Monitored by independent data monitoring committee, not named, not 30 part of sponsor or investigators.” 31 Selected comments for access to data Miscellaneous: “Not the ENTIRE dataset, but much of it and we were encouraged to ask for specific 32 analyses.”; 33 “I had access to all the data but did not have access to the database. E.g., all analyses conducted in 34 sponsor's dataset. Any analysis that we wanted was done.” 35 Selected comments for using data access 36 Miscellaneous: “I anticipate some 30 high quality papers will be published using this data set. 12 already 37 published, 4 submitted and 10 currently in preparation” 38 Selected comments for drafting the manuscript 39 Academics writing intro and discussion: “I drafted the introduction and discussion sections, while the 40 company scientific writers drafted the methodology and results sections.” 41 Selected comments for how disagreements with funder were handled by the funder 42 Miscellaneous: “Want to postpone and change the wording” 43 Selected comments from additional comments 44 Miscellaneous: “Further report not approvable or not approved by the sponsor” Typos have been corrected. Comments from different authors are separated by “;”. Words have been 45 replaced by pseudonyms to ensure the anonymity of the author. 46 47
48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 22 of 42
1 2 3 4 TABLE 1 5 CHARACTERISTICS OF THE 200 INDUSTRY-FUNDED TRIALS 6 7 Journal n % 8 New England Journal of Medicine 106 53% 9 The Lancet 62 31% 10 11 JAMA Confidential: For Review25 13% Only 12 Annals of Internal Medicine 4 2% 13 JAMA Internal Medicine 2 1% 14 BMJ 1 1% 15 Intervention 16 17 Drug 165 83% 18 Device 26 13% 19 Vaccine 9 5% 20 Comparator 21 Active treatment 79 40% 22 Multiple arms (active treatment and placebo) 27 14% 23 24 Placebo or no additional treatment 94 47% 25 Authorship 26 Academic and industry funder authors* 173 87% 27 Academic and CRO authors 5 3% 28 Solely academic authors 22 11% 29
30 31 1st author academic 196 98% 32 1st author funder 4 2% 33 Last author academic 154 77% 34 Last author funder 42 21% 35 Last author CRO 2 1% 36 Last author other** 2 1% 37 38 Corresponding author academic 192 96% 39 Corresponding author funder 8 4% 40 Lead academic author's reported conflicts of interest 41 Conflict(s) of interest with funder*** 165 83% 42 Conflict(s) of interest with other company 14 7% 43 44 No conflicts of interest 21 11% 45 CRO: –contract research organisation. Due to rounding the percentages may not add up to 100% 46 *Some trials also had CRO employed co authors 47 **Other refers to an author employed by an industry company other than the industry funder and one author 48 where it was unclear if the affiliation was a CRO or private clinic. 49 ***Those who had conflicts of interest with the funder could also have conflicts of interest with other industry 50 companies. 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 23 of 42 BMJ
1 2 3 4 TABLE 2 5 SURVEY REPORTED EXPERIENCE AMONG THE LEAD ACADEMIC AUTHORS 6 n=80* Do not know or 7 Yes No Not Available 8 9 Trial agreement n % n % n % 10 Signed trial agreement with industry Confidential: 63For 79% Review 16 20% Only 1 1% 11 funder 12 Signed trial agreement included a 13 46 73% 12 19% 5 8% publication agreement** 14 Signed trial agreement included 15 33 52% 21 33% 9 14% 16 presentation agreement** Signed trial agreement included 17 39 62% 19 30% 5 8% 18 confidentiality agreement** 19 Data access 20 Had access to the entire trial data 63 79% 9 11% 8 10% 21 set 22 Access used by those with access 23 56 89% 6 10% 1 2% to entire data*** 24 25 Problems 26 Delay in publication due to funder 3 4% 70 88% 7 9% 27 Disagreements with funder 9 11% 64 80% 7 9% 28 Collaboration 29 Already collaborating with 30 67 84% 1 1% 12 15% funder/would in the future 31 32 *7 authors only provided responses to some of the questions **n=63 Question was only available to those who answered yes to signing an agreement with industry funder 33 ***n=63 Question was only available to those who answered yes to data access 34 Due to rounding the percentages may not add up to 100%. 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 24 of 42
1 2 3 4 SUPPLEMENTAL MATERIAL 5 6 Contents 7 8 • Appendix 1 Data extraction categories 9 • Appendix 2 Survey questions 10 • Confidential: Appendix 3 Reason for declining to Forparticipate Review Only 11 • Appendix 4 Characteristics of the trials stratified by response type 12 • Appendix 5 Survey responses 13 14 • Appendix 6 Selected comments to the survey questions 15 • Appendix 7 Subgroup analyses 16 17 18 19 Appendix 1. Data extraction categories 20 21 • Name of first author 22 • Name of lead academic author 23 • Institutional address and email of lead academic author 24 • Categorisation of first, last and corresponding author as academic, funder or CRO 25 • Name of trial(s) 26 • Journal name 27 • Number of authors 28 • Number of authors involved in the statistical analysis 29 • Vaccines, drug(s) and device(s) used in trial 30 • Publication date 31 • Trial registration in a WHO approved registry (e.g. clinicaltrials.gov) 32 • Name(s) of trial industry funder(s) 33 • Verbatim extraction of author contributions statement 34 • Verbatim extraction of “Role of the funder statement”? 35 • Funder employee co-author(s) (yes/no)? 36 • Academic authors’ role in trial in relation to: 37 o study design 38 o study conduct 39 o study analysis 40 o study reporting 41 • Funder’s role in trial in relation to: 42 o study design 43 o study conduct 44 o study analysis 45 o study reporting 46 • Any use of a contract research organisation and their role in relation to: 47 o study design 48 o study conduct 49 o study analysis 50 o study reporting 51 • Lead academic author’s disclosed conflicts of interest 52 • Verbatim extraction of authors’ access to data and type of data 53 • Description of confidentiality agreements 54
55
56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 25 of 42 BMJ
1 2 3 4 Appendix 2. Survey questions 5 6 1. Contract 7 a. Was any contract(s) or agreement(s) signed between you or other academic 8 investigators and the trial funder (e.g. an investigator or publication agreement)? 9 (If yes these questions appeared): 10 i. Did it include a publication agreement (e.g. any trial publication needs 11 Confidential:approval from the funder For prior toReview submission)? Only 12 ii. Did it include a presentation agreement (e.g. any trial presentation needs 13 approval from the funder before presentation)? 14 iii. Did it include a confidentiality agreement (e.g. study results or protocol 15 information may only be shared with third party after approval from the 16 funder)? 17 iv. Did it include other types of agreements, please describe (blank text box)? 18 19 2. Benefits of collaboration with the trial funder 20 a. Please describe any benefits to your collaboration? 21 b. Would you collaborate with this funder again? 22 23 3. Trial design 24 a. Who decided what the comparator treatment should be (e.g. choice of active versus 25 placebo comparator, type and dose of comparator drug or comparator device 26 used)? 27 b. Who decided which outcomes to measure in the trial? 28 c. Who had the final say with regard to trial design? 29 30 4. Data analysis 31 a. Who performed the actual statistical analysis of the trial data i.e. using statistical 32 analysis software? 33 b. Did you personally have access to the entire dataset? 34 (If yes this question appeared) i. Did you actively use this access? 35
36 5. Manuscript 37 a. Who wrote the draft manuscript? 38 b. Who made the final decision on the content in the published manuscript (For 39 example, which outcomes to report or how data should be interpreted)? 40
41 6. Collaboration with funder 42 a. Was there any delay of publications due to funder? 43 b. Were there any disagreements between you as academic investigator(s) and the 44 funder concerning design, analysis, reporting of outcomes and/or writing the 45 publication 46 (If yes this question appeared) 47 i. Please describe how these disagreements were managed by the funder? 48
49 7. If you have any additional comments you find relevant for this survey, please describe them 50 here. 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 26 of 42
1 2 3 4 Appendix 3. Reasons for declining to participate 5 6 7 APPENDIX 3 TABLE 1 8 REASONS PROVIDED VIA EMAIL FOR NOT PARTICIPATING IN THE SURVEY 9 10 n 11 Not interestedConfidential: in participating For Review Only5 12 Lack of time 4 13 Do not wish to collaborate with the Cochrane Collaboration/The Nordic Cochrane 14 Centre 2 15 Do not have the information requested/unable to help 2 16 Concerns with objectivity of the survey 1 17 Contact funder instead 1 18 Impossible I have joined industry 1 19 20 21 Appendix 4. Characteristics of the trials stratified by survey response type 22 23 24 APPENDIX 4 TABLE 1 25 CHARACTERISTICS OF THE TRIALS STRATIFIED BY SURVEY RESPONSE TYPE Accessed survey 26 Declined Non- Responders* without participation responders** 27 responding 28 Authorship n=80 % n=10 % n=16 % n=91 % 29 Median number of authors (range) 18 (1-48) 19 (10-36) 16 (5-35) 19 (6-40) 30 Academic and industry funder authors 68 85% 8 80% 15 94% 80 88% 31 32 Solely academic authors 12 15% 2 20% 1 6% 11 12% 33 Corresponding author academic 77 96% 10 100% 16 100% 86 95% 34 Corresponding author funder 3 4% 0 0% 0 0% 5 5% 35 Comparator 36 Active treatment 33 41% 5 50% 6 38% 34 37% 37 Multiple arms (active treatment and placebo) 9 11% 1 10% 1 6% 16 18% 38 Placebo or no additional treatment 38 48% 4 40% 9 56% 41 45% 39 Role of funder 40 Funder involved in design 66 83% 6 60% 16 100% 79 87% 41 Funder involved in conduct 58 73% 7 70% 13 81% 66 73% 42 Funder involved in data analysis 53 66% 8 80% 14 88% 70 77% 43 Funder involved in reporting 68 85% 8 80% 16 100% 74 81% 44 CRO involved in reporting 43 54% 7 70% 14 88% 55 60% 45 Lead academic author's reported conflicts of interest 46 Conflict(s) of interest with funder*** 58 73% 10 100% 14 88% 78 86% 47 Conflict(s) of interest with other company 11 14% 0 0% 1 6% 4 4% 48 49 No conflict of interest 11 14% 0 0% 1 6% 9 10% CRO: contract research organisation. 50 Due to rounding the percentages may not add up to 100% 51 *7 authors only provided responses to some of the questions, 1 of them later emailed to say he did not have 52 time to complete the survey, and he has been counted under Responders in this table 53 **3 unreachable authors’ trials were not included in the non-responders. 54 ***Those who had conflicts of interest with the funder could also have conflicts of interest with another 55 industry company. 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 27 of 42 BMJ
1 2 3 4 Appendix 5. Survey responses of lead academic authors 5 6 APPENDIX 5 TABLE 1 7 SURVEY REPORTED EXPERIENCE AMONG THE LEAD ACADEMIC AUTHORS 8 9 10 n=80* Confidential:Academic ForFunder** ReviewCRO** Regulator Only Other*** 11
12 n % n % n % n % n % 13 Choice of comparator 26 33% 12 15% 0 0% 6 8% 0 0% 14 15 Choice of outcomes 4 5% 23 29% 0 0% 5 6% 1 1% 16 Final say in design 29 36% 17 21% 0 0% 3 4% 0 0% 17 Conducted statistical analysis 21 26% 28 35% 7 9% 0 0% 1 1% 18 Drafted manuscript 53 66% 8 10% 0 0% 0 0% 1 1% 19 Final say on the published manuscript 52 65% 0 0% 0 0% 0 0% 1 1% 20 Academic, Academic Funder 21 funder Academic Funder and and Do not and/or and CRO and CRO 22 n=80* regulator regulator know/Not CRO collaborati collaborati collaborati collaborati available 23 collaborati on on on on 24 on 25 n % n % n % n % n % n % 26 Choice of comparator 23 29% 1 1% 0 0% 4 5% 2 3% 6 8% 27 Choice of outcomes 35 44% 0 0% 0 0% 3 4% 3 4% 6 8% 28 29 Final say in design 24 30% 0 0% 0 0% 0 0% 2 3% 5 6% 30 Conducted statistical analysis 11 14% 0 0% 5 6% 0 0% 0 0% 7 9% 31 Drafted manuscript 8 10% 3 4% 0 0% 0 0% 0 0% 7 9% 32 Final say on the published manuscript 19 24% 1 1% 0 0% 0 0% 0 0% 7 9% 33 *7 authors only provided responses to some of the questions 34 **Funder and CRO also includes unacknowledged persons for the conducted statistical analysis and drafting 35 of manuscript. In 7 trials the statistical analysis was conducted by a funder or CRO employee who was not a 36 named author or contributor (3 of the 7 trials had unacknowledged funder employees, 3 trials had 37 unacknowledged CRO employees and 1 trial had both an unacknowledged funder and unacknowledged 38 CRO employee conduct the statistical analysis). In 5 trials the manuscript was drafted by a funder or CRO 39 employee who was not a named author or contributor (4 of the 5 trials had an unacknowledged funder employee and 1 trial had an academic and an unacknowledged CRO draft the manuscript). Two trials had 40 both an unacknowledged funder or CRO employee conduct the statistical analysis and an unacknowledged 41 funder or CRO employee draft the manuscript. Thus, lead academic authors of 10 trials reported 42 contributions to statistical analysis and/or drafting of the manuscript from unacknowledged persons. 43 ***Other refers to one trial where it was unclear who had chosen outcomes and conducted statistical analysis 44 and one where the academic did not know who drafted the manuscript and one where the academic found 45 that the journal had the final say on the published manuscript. 46 Due to rounding the percentages may not add up to 100%. 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 28 of 42
1 2 3 4 Appendix 6. Comments to the survey questions 5 6 Different author comments have been separated by “;”. Words have been replaced by 7 “pseudonyms” to anonymise the author. Typos have been corrected. 8 9 Comments from Benefits of collaborating with the trial industry funder 10 Generating new research: “They were willing to do a global trial of an orphan disease that no other 11 sourceConfidential: would have EVER funded.”; For Review Only 12 “Chance to move the field forward and to develop medicine based on prior lab discoveries” 13 Infrastructure and management: “Efficient operational management of multi-center trial”; 14 “Much faster at helping address quality concerns and organizing than government” 15 Funding: “funding (much better than with governmental funding)”; 16 “Fully funded trial with additional funding for translational research to be done by co-operative 17 group”; 18 “Study would not have been possible without funding which came as a donation, no strings 19 attached.”; 20 “Funding (in this case, there would have been no funding to conduct the trial otherwise)” 21 Publication: “The Journal X paper! And more to come.”; 22 “Opportunity to author papers of important study results in high impact journals” 23 Personal benefits: “Being PI of a positive trial is always a benefit in CV and recognition”; 24 “I was able to work closely with the study team to write and revise the protocol. We had a lot of 25 discussions and I generally found that they listened to the advice offered by the trial steering 26 committee. I was able to review the data prior to drafting the manuscript and we had a few calls to 27 discuss the implications of the data and how to focus the paper. I found their statistical team 28 generally helpful - at times they pushed back because I asked for a lot of additional analyses but 29 they were generally accommodating. After the first draft was written, we definitely had some back 30 and forth on messaging but ultimately, I was very comfortable with the final product and found that 31 they listened to my point of view. Aside from the process itself, there are 'academic' benefits in the 32 sense that this type of study is highly regarded by our institution and brings some notoriety to our 33 institution. That being said, I would not object to all industry-sponsored studies being listed as 'the 34 Trial X Investigators' or something similar - ultimately, despite our input, these are studies done by 35 the company.”; 36 “consulting fee, publication of research” 37 Miscellaneous: “very fruitful and pragmatic collaboration with funder, allowing to transparently enrol 38 patients in our country, and very transparent discussions of results, that is what I call "benefits"”; 39 “Not really any other than data analysis help”; 40 “Industry funded the trial at substantial cost. They organized the sites. They created most of the 41 study materials. Contractors did most of the work, contributed to design, analysis and reporting. 42 Industry could do it themselves I assume. I do think the partnership brings value to both parties 43 and patients.”; 44 “Input into protocol, input to data interpretation, development of scientific expertise in clinical trials 45 and in the subject are of research, working with other key international leaders in the field, scientific 46 excellence of the study sponsor, well-resourced study.”; 47 “The major benefits of industry collaboration for this investigator designed, led and run study were 48 the provision of funding, labelled IMP and placebo, and logistical support.”; 49 “Transparency, faster results and execution, excellent funding.”; 50 “You can answer questions that academic or government funders cannot do. This mainly due to 51 the scale of research funding needed and the risks taken. / There is better oversight of industry 52 funded research. Much stricter auditing, monitoring and reporting of events etc.”; 53 “funding provided. Help in data analysis and publication provided” 54 No benefits: “No real benefits....”; 55 “Nothing” 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 29 of 42 BMJ
1 2 3 4 Comments from trial agreement between academics and funder 5 Miscellaneous: “Very thorough, well prepared With mutual respect”; 6 “Non-disclosure Agreement / Clinical Trial Agreement w/ Payment schedule” 7 Comments from publication agreement 8 Review and comment: “the trial publication can be reviewed and comments provided. / but they do 9 not get "approval" rights.”; 10 “2-week period to allow for input, which the executive committee (independent of sponsor) could 11 chooseConfidential: to ignore.”; For Review Only 12 “Not “approval” but rather an opportunity for them to make comments (which we could either take 13 or reject)” 14 Miscellaneous: “Yes and aimed at journals With high impact factor”; 15 “the contract explicitly stated that sponsor approval was NOT required for publication.” 16 Comments for presentation agreement 17 Review and comment: “Stipulated right of review by sponsor”; 18 “they would be sent for review by sponsor but we have final say”; 19 “I underline that there was an agreement, but in no way some kind of censorship.” 20 Comments for confidentiality agreement 21 Confidential until publication: “Until approval or public disclosure of the information.”; 22 “No for results which are already in the public domain, obviously, but yes for all other unpublished 23 data”; 24 “Prior to presentation & simultaneous publication, the data was 100% embargoed for all parties” 25 Miscellaneous: “At some point I did sign confidentiality agreement. At beginning of trial and then 26 before reviewing final data.“; 27 “after publication.....”; 28 “non-public information could not be shared with a 3rd party - for obvious reasons”; 29 “the agreement stated that sponsor could request up to 30 days confidentiality if necessary for 30 patent protection.” 31 Comments for other types of agreement 32 Miscellaneous: “The contract required the trial database to be transferred to the academic lead 33 author's institution for statistical analysis and publication”; 34 “adhere to good practices”; 35 “Intellectual property clauses. Scope of work. Indemnification. Termination conditions. Many 36 additional aspects...” 37 Comments for future collaboration with the funder 38 Miscellaneous: “I have done so for over 30 years”; 39 “Such collaborations are essential (provided independent data analysis, publication and 40 presentation is guaranteed) as trials of this size are rarely supported by charities or governmental 41 institutions.” 42 Comments for choice of comparator 43 Academics: “The entire trial design was not influenced by the funder or any other party” 44 Funder: “As mentioned, this is a substudy of Drug X in the context of a RCT. The funder designed 45 the study. 46 Collaboration: “The company has to obviously agree with the study design. They fund the trial.”; 47 “There was a discussion between the pharma key people and the clinical investigator team about 48 the best study design.”; 49 “Funder, FDA, academic steering committee”; 50 “Me and academic steering committee in discussions with industry partner. Industry partner has 51 final say”; 52 “Arduous, iterative process with input from FDA, academic steering committee, funder, and site 53 investigators.”; 54 “Research and development team of the sponsor. Input was obtained from study investigators.” 55 Regulatory body: “In accordance with FDA and EMA”; 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 30 of 42
1 2 3 4 “Having a placebo control in a double blinded trial design was recommended by the FDA” 5 Miscellaneous: “The sponsor did not want a placebo arm in the trial. We insisted upon it for safety 6 evaluation and they ultimately agreed to a placebo arm with deferred treatment.”; 7 “The Steering committee specifically designed the trial - and we had an active comparator and the 8 trial was neutral compared to the comparator. If we would have chosen Drug X (one might assume 9 with industry or even guidelines that would be reasonable - maybe different finding) - but the SC 10 chose the comparator” 11 CommentsConfidential: for choice of outcomes For Review Only 12 Academic: “A steering committee of academic advisors, including me.“; 13 “Again, the funder had no influence on our primary and secondary outcomes.” 14 Regulatory body: “FDA” 15 Collaboration: “Collaboration between funder, academics and of course the funder incorporated 16 suggestions from regulatory bodies”; 17 “Funder did primarily but they gathered many opinions including investigators and FDA/EMA”; 18 “This was collaboration between academic investigators, regulators (Phase IV commitment) and 19 industry sponsor” 20 Miscellaneous: “Entire trial was designed by the steering committee - this is actually less 21 involvement than government sources often have in designing trials where they are hyper- 22 focussed on cost” 23 Comments for final say in study design 24 Academics: “The principal investigator, which was the head of the research team.” 25 Funder: “We advised the sponsor study team (extensively) and I would generally say that they 26 listened, but they did not 'require' sign-off from the steering committee before finalizing the protocol 27 so in that sense, they had final say.”; 28 “Again there was discussion but final saying by industry” 29 Regulatory body: “FDA of course- it was a regulatory study so they had to approve any study 30 design” 31 Collaboration: “Again, making use of the funder's experience together”; 32 “Funder and FDA”; 33 “It was a combined decision of investigators, sponsor and finally FDA.” 34 Miscellaneous: “On most aspects, academic steering committee members and funder were in 35 agreement, so question of final say did not arise. On several aspects, FDA had final say, placing 36 requirements that would not have been selected by the academic steering committee members or 37 the funder. On other aspects, the community had say by having specialty societies issue 38 guidances. These were sometimes poorly selected directives, but the funder felt it important to 39 comply with international specialty society recommendations.” 40 Comments for statistical analysis 41 Academics: “A statistician and I performed the analysis. No influence or contact with the funder in 42 the whole process.” 43 Funder: “Biostatistical team of the funder”; 44 “Lots of analyses by a number of funder statisticians” 45 CRO: “Independent statistical bureau”; 46 “The CRO agency was hired for this purpose.” 47 Collaboration: “Jointly between Funder and Academic Investigators”; 48 “Statistical consultant, along with study team and appropriate funder personnel” 49 Miscellaneous: “Independent biostatistician contracted by study sponsor and in-house industry 50 biostatisticians.”; 51 “We analysed with independent academic and with the funder statisticians. Sometimes we have to 52 rely on only the funder statisticians, which is less satisfactory. However SAPs and protocols are 53 submitted to regulatory and editors before analysis and publication”; 54 “Most of the analysis was done by the sponsor statistician who is listed as an author but the 55 aggregated data were available to us (steering committee) and individual patient-level data was 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 31 of 42 BMJ
1 2 3 4 available upon request. We were able to do additional analyses using these data - some of which 5 ultimately ended up in the paper.”; 6 “Stats team employed by company. Monitored by independent data monitoring committee, not 7 named, not part of sponsor or investigators.” 8 Comments for access to data 9 Miscellaneous: “It varies from study to study”; 10 “Although for person-level data, access was provided in response to questions. I had an 11 aggregatedConfidential: data set, not person-level data.”; For Review Only 12 “Not the ENTIRE dataset, but much of it and we were encouraged to ask for specific analyses.”; 13 “I had access to all the data but did not have access to the database. E.g., all analyses conducted 14 in sponsor's dataset. Any analysis that we wanted was done.”; 15 “...in theory yes, but I didn't review all data”; 16 “If I ask.” 17 Comments for using data access 18 Miscellaneous: “The dataset was open for the research team and locked or the rest of the 19 members“; 20 “I anticipate some 30 high-quality papers will be published using this data set. 12 already 21 published, 4 submitted and 10 currently in preparation”; 22 “Not yet, but it is available to me for secondary analyses”; 23 “Huge database housed at our institution. Could not possibly review every page, but all of the key 24 outcomes reviewed carefully”; 25 “I personally requested the independent statistician perform many analyses of the entire dataset, 26 using shell tables that I designed. Every analysis I requested was performed.“ 27 Comments for drafting the manuscript 28 Intro and discussion by academics methods and results by funder: “I drafted the introduction and 29 discussion sections, while the company scientific writers drafted the methodology and results 30 sections.”; 31 “It was really a collaboration. The methods and initial draft of the results was written by a medical 32 writer who works for the sponsor (and is acknowledged) but the intro and discussion were written 33 by me and the last author on the paper. We then provided comment and revisions to the sections 34 written by the sponsor and with back and forth iterations, we came to a final draft that was 35 circulated to the other authors.” 36 Miscellaneous: “I did and it was reviewed by the funder's scientific collaborators. Disagreements 37 regarding data interpretation were hotly discussed but the academic point of view prevailed”; 38 “In collaboration with the co-authors of the funder”; 39 “This was an unusually coherent scientific partnership”; 40 “Myself and an academic colleague”; 41 “I wrote the first draft together with one funding representative and one other academic 42 collaborator. The other authors commented on that and subsequent drafts.” 43 Comments for final say on published manuscript 44 Journal: “Often also unfortunately - the journal has a larger and larger role”; 45 “The journal!”; 46 ”In part, these decisions were also impacted upon by Journal X whose editorial policies regarding 47 manuscript length and number of figures and tables forced us to consolidate. Ultimately, it was my 48 responsibility to do this along with my academic colleagues and the statistician” 49 Academics: “Absolute no influence of the funder. We kept them out during the process.” 50 Miscellaneous: “Company had review privilege but investigators had final say”; 51 “There is input from the sponsors but the final decisions is with all authors, some who may work for 52 funders”; 53 “Primarily the Investigators/authors. Some guidance from funder.”; 54 “The funder and the authors did together....”; 55 “Investigator in agreement with sponsor” 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 32 of 42
1 2 3 4 Comments for delay in publication 5 Miscellaneous: “There was a delay in publication, which was related to internal data check.”; 6 “Actually, less delay”; 7 “It varies from trial to trial”; 8 “Despite the stipulations in the contract, the manuscript was not sent to the funder prior to 9 submission. This was considered unnecessary by both sides (despite the negative result, i.e. the 10 medication is completely ineffective)” 11 CommentsConfidential: for disagreements with funder For Review Only 12 Miscellaneous: “The sponsors were originally reluctant to run the study, but we won over by the 13 academic advisors”; 14 “The interpretation criteria stated in the protocol were partially contradicted by the data forcing to 15 present data according to protocol and new criteria. This was actually requested by the publisher.”; 16 “When an external trial resulted positive, we needed to decide whether to place enrolment in the 17 current trial on hold until a slightly earlier than planned interim analysis could be performed, or to 18 continue enrolling. The funder would have preferred to continue enrolling. The academic steering 19 committee voted 4 to 2 place the study on hold. The funder respected this decision and enrolment 20 was placed on hold.” 21 Minor disagreements: “Spirited discussion but agreement usually prevails”; 22 “Small differences between PI/statistician and funder. The final version was exactly the version as 23 proposed by PI/Statistician”; 24 “The 'disagreements' were minor. Mostly on the points to emphasize in the discussion. Ultimately 25 they accepted very close to our original version.”; 26 “Minimal. There were some analyses some investigators wanted to do but truly were beyond the 27 scope of the primary goals of the trial.” 28 Comments for how disagreements with funder were handled by the funder 29 Miscellaneous: “Want to postpone and change the wording”; 30 “Back and forth emails and teleconferences. As noted, ultimately, they agreed on very close to our 31 original version.”; 32 “We initially communicated about the disagreements via email, but with continued analysis and 33 discussion, specific teleconferences were set up to improve communication efforts.”; 34 “The journal and I ended up determining the outcome”; 35 “Other than the funder author, who gave appropriate scientific input, they were handled at arm's 36 length.” 37 Comments from additional comments 38 Miscellaneous: “For this research, academic independence was of importance for the academic 39 staff. That is what was agreed with the funder, whether they liked it or not. We could provide the 40 funder an academic setting, with highly skilled personnel to conduct the study. And that was what 41 they wanted too. So for both groups it was a win-win.”; 42 “The study was somewhat unusual for pharma phase III studies, with greater input from the 43 academic investigators, including study design, manuscript preparation, analyses etc.”; 44 “It is incredibly important that you distinguish between academically-led and performed trials, such 45 as those conducted by my unit, and industry-led, conducted and analysed trials that may well have 46 a degree of academic oversight but with most of not all of the rights remaining with the company”; 47 “We had to manage possible conflicts of interest at our sites carefully.”; 48 “Discussions with the funder had a high intellectual and scientific level”; 49 “The study and its publication brought prestige to Society X. Investigators are very keen to join 50 hands for more studies now.”; 51 “Would have preferred to have the whole dataset. Got quite a lot but not everything”; 52 “I found participating in this project to be a quite positive experience. I did not feel any pressure 53 from the funder, believe I functioned in a totally independent manner, was able to obtain 54 constructive feedback from my two primary academic colleagues as well as the other academic 55 authors, and received no payment or grant for this project other than travel expenses for one 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 33 of 42 BMJ
1 2 3 4 planning meeting.”; 5 “In principle this type of partnership is vital to move the field forward. Rules for collaboration need 6 to be set in advance, ideally on a contractual basis taking academic freedom as well as the needs 7 of the industry into account, which can be tough, especially in the light of patent rights.”; 8 “Further report not approvable or not approved by the sponsor” 9 10 11 Confidential: For Review Only 12 13 Appendix 7. Subgroup analyses 14 15 Funder Problems stratified by Lead Academic’s Conflicts of Interest 16 17 The nine academic authors with disclosed conflicts of interest (COIs) with an industry company 18 other than the funder were not included in this subgroup analysis. 19 20 Responders with no COIs (n = 11) 21 One reported there was a delay in publication due to the funder doing an internal data check. Eight 22 23 reported no delays in publication and two did not provide a response. 24 25 Similarly only one responder reported to have had minor disagreements about wording with the 26 industry funder. Eight reported no disagreements but one added that the funder was more 27 interested in the secondary outcome. Two did not provide a response. 28 29 Responders with COIs with the funder (n = 60) 30 31 Two reported there was a delay in publication due to the funder with one reporting that this was 32 due to internal approval and one author responding that delay in publication varies from trial to trial. 33 Fifty-four reported no delays and 4 did not provide a response. 34 35 Seven reported they had disagreements with the industry funder. Most of the authors added that 36 the disagreements were minor. One reported that the funder was reluctant to run the study and 37 one reported disagreements between the PI/statistician and the funder. Forty-nine reported no 38 disagreements, but one added they had spirited discussions, but agreement usually prevailed. 39 Four did not provide a response. 40 41
42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 34 of 42
1 2 3 4 APPENDIX 7. TABLE 1A 5 CHARACTERISTICS OF THE 200 INDUSTRY-FUNDED 6 TRIALS BY INTERVENTION TYPE 7 8 Drug Device Vaccine 9 n % n % n % 10 Journal 11 New EnglandConfidential: Journal of Medicine For92 Review 56% 9 35%Only 5 56% 12 The Lancet 49 30% 10 38% 3 33% 13 JAMA 17 10% 7 27% 1 11% 14 Annals of Internal Medicine 4 2% 0 0% 0 0% 15 JAMA Internal Medicine 2 1% 0 0% 0 0% 16 17 BMJ 1 1% 0 0% 0 0% 18 Comparator 19 Active treatment 55 33% 20 77% 4 44% 20 Multiple arms (active treatment and placebo) 27 16% 0 0% 0 0% 21 Placebo or no additional treatment 83 50% 6 23% 5 56% 22 Authorship 23 24 Median (range) 18 (1-48) 17 (6-35) 27 (17-31) 25 Academic and industry funder authors 151 92% 13 50% 9 100% 26 Solely academic authors 14 8% 13 50% 0 0% 27 28 1st author academic 162 98% 26 100% 8 89% 29 30 1st author funder 3 2% 0 0% 1 11% 31 Last author academic 127 77% 23 88% 4 44% 32 Last author funder 35 21% 2 8% 5 56% 33 Last author CRO 1 1% 1 4% 0 0% 34 Last author other* 2 1% 0 0% 0 0% 35 Corresponding author academic 160 97% 26 100% 6 67% 36 37 Corresponding author funder 5 3% 0 0% 3 33% 38 Lead academic author's reported conflicts of interest 39 Conflict(s) of interest with funder** 139 84% 19 73% 7 78% 40 Conflict(s) of interest with other company 12 7% 2 8% 0 0% 41 No conflict of interest 14 8% 5 19% 2 22% 42 CRO: contract research organisation. 43 Due to rounding the percentages may not add up to 100% 44 *Other refers to an author employed by an industry company other than the industry funder and one author 45 where it was unclear if the affiliation was a CRO or private clinic. 46 **Those who had conflicts of interest with the funder could also have conflicts of interest with another 47 industry company. 48 49 50
51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 35 of 42 BMJ
1 2 3 4 APPENDIX 7. TABLE 1B 5 CHARACTERISTICS OF INDEPENDENT ACADEMIC TRIALS 6 AND TRIALS WITH FUNDER INVOLVEMENT 7 Independent* Funder Involved 8 n=8 n=192 9 n % n % 10 Journal 11 New EnglandConfidential: Journal of Medicine For1 Review 13% 105 Only 55% 12 The Lancet 2 25% 60 31% 13 JAMA 3 38% 22 11% 14 Annals of Internal Medicine 1 13% 3 2% 15 JAMA Internal Medicine 1 13% 1 1% 16 17 BMJ 0 0% 1 1% 18 Comparator 19 Active treatment 4 50% 75 39% 20 Multiple arms (active treatment and placebo) 0 0% 27 14% 21 Placebo or no additional treatment 4 50% 90 47% 22 Authorship 23 24 Median (range) 13 (6-21) 19 (1-48) 25 Academic and industry funder authors 0 0% 173 90% 26 Solely academic authors 8 100% 19 10% 27 28 1st author academic 8 100% 188 98% 29 30 1st author funder 0 0% 4 2% 31 Last author academic 8 100% 146 76% 32 Last author funder 0 0% 42 22% 33 Last author CRO 0 0% 2 1% 34 Last author other** 0 0% 2 1% 35 Corresponding author academic 8 100% 184 96% 36 37 Corresponding author funder 0 0% 8 4% 38 Lead academic author's reported conflicts of interest 39 Conflict(s) of interest with funder*** 4 50% 161 84% 40 Conflict(s) of interest with other company 0 0% 14 7% 41 No conflict of interest 4 50% 17 9% 42 CRO: contract research organisation. 43 Due to rounding the percentages may not add up to 100% 44 *Independent trials defined as trials with no funder or CRO co-authors and no funder or CRO involvement in 45 the design, conduct, analysis and reporting of the trial. 46 **Other refers to an author employed by an industry company other than the industry funder and one author 47 where it was unclear if the affiliation was a CRO or private clinic. 48 ***Those who had conflicts of interest with the funder could also have conflicts of interest with another 49 industry company. 50
51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 36 of 42
1 2 3 4 APPENDIX 7. TABLE 2A 5 SURVEY REPORTED EXPERIENCE AMONG THE LEAD ACADEMIC AUTHORS STRATIFIED BY INTERVENTION TYPE 6 7 Drug n=61 Academic Funder* CRO* Regulator Other** 8 n % n % n % n % n %
9 Choice of comparator 20 33% 11 18% 0 0% 5 8% 0 0% 10 Choice of outcomes 17 28% 3 5% 0 0% 3 5% 1 2% 11 Confidential: For Review Only Final say in design 23 38% 14 23% 0 0% 1 2% 0 0% 12 13 Conducted statistical analysis 18 30% 25 41% 3 5% 0 0% 1 2% 14 Drafted manuscript 41 67% 7 11% 0 0% 0 0% 1 2% 15 Final say on the published manuscript 39 64% 0 0% 0 0% 0 0% 1 2% Academic, Academic 16 Funder and Do not funder and Drug n=61 regulator know/Not 17 and/or CRO regulator collaboration available 18 collaboration collaboration 19 n % n % n % n % 20 21 Choice of comparator 18 30% 4 7% 0 0% 3 5% 22 Choice of outcomes 29 48% 3 5% 2 3% 3 5% 23 Final say in design 18 30% 0 0% 2 3% 3 5% 24 Conducted statistical analysis 10 16% 0 0% 0 0% 4 7% 25 Drafted manuscript 8 13% 0 0% 0 0% 4 7% 26 Final say on the published manuscript 17 28% 0 0% 0 0% 4 7% 27 28 Device n=13 Academic Funder* CRO* Regulator Other** 29 30 n % n % n % n % n % 31 32 Choice of comparator 4 31% 0 0% 0 0% 1 8% 0 0% 33 Choice of outcomes 5 38% 0 0% 0 0% 1 8% 0 0% 34 Final say in design 5 38% 1 8% 0 0% 2 15% 0 0% 35 Conducted statistical analysis 3 23% 2 15% 3 23% 0 0% 0 0% 36 Drafted manuscript 10 77% 0 0% 0 0% 0 0% 0 0% 37 Final say on the published manuscript 10 77% 0 0% 0 0% 0 0% 0 0% 38 Academic, Academic 39 Funder and Do not funder and Device n=13 regulator know/Not 40 and/or CRO regulator collaboration available 41 collaboration collaboration
42 43 n % n % n % n % 44 Choice of comparator 5 38% 0 0% 1 8% 2 15% 45 Choice of outcomes 5 38% 0 0% 0 0% 2 15% 46 Final say in design 3 23% 0 0% 0 0% 2 15% 47 Conducted statistical analysis 3 23% 0 0% 0 0% 2 15% 48 Drafted manuscript 1 8% 0 0% 0 0% 2 15% 49 Final say on the published manuscript 1 8% 0 0% 0 0% 2 15% 50 Vaccine n=6 Academic Funder* CRO* Regulator Other** 51 52 n % n % n % n % n % 53 Choice of comparator 2 33% 1 17% 0 0% 0 0% 0 0% 54 Choice of outcomes 1 17% 1 17% 0 0% 1 17% 0 0% 55 Final say in design 1 17% 2 33% 0 0% 0 0% 0 0% 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 37 of 42 BMJ
1 2 3 4 Conducted statistical analysis 0 0% 1 17% 1 17% 0 0% 0 0% 5 Drafted manuscript 2 33% 1 17% 1 17% 0 0% 0 0% 6 Final say on the published manuscript 3 50% 0 0% 0 0% 0 0% 0 0% 7 Academic, Academic Funder and Do not 8 funder and Vaccine n=6 regulator know/Not and/or CRO regulator 9 collaboration available collaboration collaboration 10 Confidential: For Review Only 11 n % n % n % n % 12 13 Choice of comparator 1 17% 0 0% 1 17% 1 17% 14 Choice of outcomes 1 17% 0 0% 1 17% 1 17% 15 Final say in design 3 50% 0 0% 0 0% 0 0% 16 Conducted statistical analysis 3 50% 0 0% 0 0% 1 17% 17 Drafted manuscript 1 17% 0 0% 0 0% 1 17% 18 19 Final say on the published manuscript 2 33% 0 0% 0 0% 1 17% *Funder and CRO also includes unacknowledged persons for the conducted statistical analysis and drafting 20 of manuscript 21 **Other refers to one trial where it was unclear who had chosen outcomes and conducted statistical analysis 22 and one where the academic did not know who drafted the manuscript and one where the academic found 23 that the journal had the final say on the published manuscript. 24 Due to rounding the percentages may not add up to 100%. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 38 of 42
1 2 3 4 5 APPENDIX 7. TABLE 2B 6 SURVEY REPORTED EXPERIENCE AMONG THE LEAD ACADEMIC AUTHORS STRATIFIED AS INDEPENDENT ACADEMIC TRIALS AND TRIALS WITH FUNDER INVOLVEMENT 7 8 Independent n=4 Academic Funder* CRO* Regulator Other** 9 n % n % n % n % n % 10 11 Choice ofConfidential: comparator 4For 100% Review 0 0% 0 Only 0% 0 0% 0 0% 12 Choice of outcomes 4 100% 0 0% 0 0% 0 0% 0 0% 13 Final say in design 4 100% 0 0% 0 0% 0 0% 0 0% 14 Conducted statistical analysis 4 100% 0 0% 0 0% 0 0% 0 0% 15 Drafted manuscript 4 100% 0 0% 0 0% 0 0% 0 0% 16 Final say on the published manuscript 4 100% 0 0% 0 0% 0 0% 0 0% 17 Academic, Academic 18 funder Funder and Do not and Independent n=4 and/or CRO regulator know/Not 19 regulator collaboratio collaboration available collaboration 20 n
21 n % n % n % n % 22 Choice of comparator 0 0% 0 0% 0 0% 0 0% 23 24 Choice of outcomes 0 0% 0 0% 0 0% 0 0% 25 Final say in design 0 0% 0 0% 0 0% 0 0% 26 Conducted statistical analysis 0 0% 0 0% 0 0% 0 0% 27 Drafted manuscript 0 0% 0 0% 0 0% 0 0% 28 Final say on the published manuscript 0 0% 0 0% 0 0% 0 0% 29 30 Funder involved n=76 Academic Funder* CRO* Regulator Other** 31 32
33 n % n % n % n % n % 34 Choice of comparator 22 29% 12 16% 0 0% 6 8% 0 0% 35 Choice of outcomes 19 25% 4 5% 0 0% 5 7% 1 1% 36 Final say in design 25 33% 17 22% 0 0% 3 4% 0 0% 37 Conducted statistical analysis 17 22% 28 37% 7 9% 0 0% 1 1% 38 39 Drafted manuscript 49 64% 8 11% 0 0% 0 0% 1 1% 40 Final say on the published manuscript 48 63% 0 0% 0 0% 0 0% 1 1% 41 Academic, Academic funder Funder and Do not 42 and Funder involved n=76 and/or CRO regulator know/Not regulator 43 collaboratio collaboration available collaboration 44 n
45 n % n % n % n % 46 47 Choice of comparator 24 32% 4 5% 2 3% 6 8% 48 Choice of outcomes 35 46% 3 4% 3 4% 6 8% 49 Final say in design 24 32% 0 0% 2 3% 5 7% 50 Conducted statistical analysis 16 21% 0 0% 0 0% 7 9% 51 Drafted manuscript 11 14% 0 0% 0 0% 7 9% 52 53 Final say on the published manuscript 20 26% 0 0% 0 0% 7 9% *Funder and CRO also includes unacknowledged persons for the conducted statistical analysis and drafting 54 of manuscript 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 39 of 42 BMJ
1 2 3 4 **Other refers to one trial where it was unclear who had chosen outcomes and conducted statistical analysis 5 and one where the academic did not know who drafted the manuscript and one where the academic found 6 that the journal had the final say on the published manuscript. Due to rounding the percentages may not add up to 100%. 7 8 APPENDIX 7. TABLE 2C 9 SURVEY REPORTED EXPERIENCE AMONG THE LEAD ACADEMIC AUTHORS STRATIFIED BY COLLABORATION STATUS 10 CollaboratingConfidential: with funder/ would in the future For Review Only Other** Academic Funder** CRO** Regulator 11 n=67 * 12 n % n % n % n % n % 13 Choice of comparator 23 34% 10 15% 0 0% 6 9% 0 0% 14 Choice of outcomes 19 28% 3 4% 0 0% 5 7% 1 1% 15 Final say in design 25 37% 15 22% 0 0% 3 4% 0 0% 16 Conducted statistical analysis 19 28% 26 39% 4 6% 0 0% 1 1% 17 Drafted manuscript 48 72% 6 9% 0 0% 0 0% 1 1% 18 Final say on the published manuscript 47 70% 0 0% 0 0% 0 0% 1 1% 19 Academic, Academic Funder and 20 funder and Do not Collaborating with funder/ would in the future regulator and/or CRO regulator know/Not 21 n=67 collaboratio collaboratio collaboratio available n 22 n n 23 n % n % n % n % 24 Choice of comparator 23 34% 4 6% 1 1% 0 0% 25 Choice of outcomes 33 49% 3 4% 3 4% 0 0% 26 Final say in design 22 33% 0 0% 2 3% 0 0% 27 Conducted statistical analysis 16 24% 0 0% 0 0% 1 1% 28 Drafted manuscript 11 16% 0 0% 0 0% 1 1% 29 Final say on the published manuscript 18 27% 0 0% 0 0% 1 1% 30 Would not collaborate with funder again or does Other** 31 Academic Funder** CRO** Regulator 32 not know n=8* *
33 n % n % n % n % n % 34 Choice of comparator 3 38% 2 25% 0 0% 0 0% 0 0% 35 36 Choice of outcomes 4 50% 1 13% 0 0% 0 0% 0 0% 37 Final say in design 4 50% 2 25% 0 0% 0 0% 0 0% 38 Conducted statistical analysis 2 25% 2 25% 3 38% 0 0% 0 0% 39 Drafted manuscript 5 63% 2 25% 0 0% 0 0% 0 0% 40 Final say on the published manuscript 5 63% 0 0% 0 0% 0 0% 0 0% Academic, Academic 41 Funder and funder and Do not Would not collaborate with funder again or does regulator 42 and/or CRO regulator know/Not not know n=8* collaboratio collaboratio collaboratio available 43 n 44 n n 45 n % n % n % n % 46 Choice of comparator 1 13% 0 0% 1 13% 1 13% 47 Choice of outcomes 2 25% 0 0% 0 0% 1 13% 48 Final say in design 1 13% 0 0% 0 0% 1 13% 49 Conducted statistical analysis 0 0% 0 0% 0 0% 1 13% 50 Drafted manuscript 0 0% 0 0% 0 0% 1 13% 51 Final say on the published manuscript 2 25% 0 0% 0 0% 1 13% 52 *1 of the 8 authors would not collaborate with funder again, the remaining 7 did not know. 53 **Funder and CRO also includes unacknowledged persons for the conducted statistical analysis and drafting 54 of manuscript 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 40 of 42
1 2 3 4 ***Other refers to one trial where it was unclear who had chosen outcomes and conducted statistical analysis 5 and one where the academic did not know who drafted the manuscript and one where the academic found 6 that the journal had the final say on the published manuscript. Due to rounding the percentages may not add up to 100%. 7 8 9 APPENDIX 7. TABLE 3A 10 11 SURVEYConfidential: REPORTED EXPERIENCE AMONG THE LEAD For ACADEMIC ReviewAUTHORS STRATIFIED BY Only INTERVENTION TYPE 12 Drug n=61 13 Do not know or Yes No 14 Not Available 15 16 n % n % n % 17 Signed trial agreement with industry funder 50 82% 11 18% 0 0% 18 36 72% 10 20% 3 6% 19 Signed trial agreement included a publication agreement* 26 52% 18 36% 5 10% 20 Signed trial agreement included presentation agreement* 21 28 56% 17 34% 4 8% 22 Signed trial agreement included confidentiality agreement* 23 24 Had access to the entire trial data set 47 77% 9 15% 5 8% 25 26 42 89% 4 9% 1 2% Access used by those with access to entire data* 27 28 2 3% 55 90% 4 7% 29 Delay in publication due to funder 4 7% 53 87% 4 7% 30 Disagreements with funder 31 Device n=13 32 Do not know or 33 Yes No 34 Not Available 35 n % n % n % 36 8 62% 5 38% 0 0% Signed trial agreement with industry funder 37 5 63% 2 25% 1 13% 38 Signed trial agreement included a publication agreement* 39 3 38% 2 25% 3 38% 40 Signed trial agreement included presentation agreement* 41 7 88% 1 13% 0 0% Signed trial agreement included confidentiality agreement* 42 43 Had access to the entire trial data set 11 85% 0 0% 2 15% 44 9 82% 2 18% 0 0% 45 Access used by those with access to entire data* 46 47 Delay in publication due to funder 1 8% 10 77% 2 15% 48 Disagreements with funder 4 31% 6 46% 2 15%
49 Vaccine n=6 50 Do not know or Yes No 51 Not Available 52 53 n % n % n % 5 83% 0 0% 1 17% 54 Signed trial agreement with industry funder 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj Page 41 of 42 BMJ
1 2 3 4 5 100% 0 0% 0 0% 5 Signed trial agreement included a publication agreement* 4 80% 1 20% 0 0% 6 Signed trial agreement included presentation agreement* 7 4 80% 1 20% 0 0% 8 Signed trial agreement included confidentiality agreement* 9 10 Had accessConfidential: to the entire trial data set For5 Review 83% 0 0%Only 1 17% 11 Access used by those with access to entire data* 5 100% 0 0% 0 0% 12 0 0% 5 83% 1 17% 13 Delay in publication due to funder 14 0 0% 5 83% 1 17% 15 Disagreements with funder 16 Due to rounding the percentages may not add up to 100% 17 *n=50, n=8 and n=5 for drug, device and vaccine trials, respectively. Question was only available to those 18 who answered yes to signing an agreement with industry funder **n=47, n=11 and n=5 for drug, device and vaccine trials, respectively. Question was only available to those 19 who answered yes to data access. 20 21 APPENDIX 7. TABLE 3B 22 SURVEY REPORTED EXPERIENCE AMONG THE LEAD ACADEMIC AUTHORS 23 STRATIFIED AS INDEPENDENT ACADEMIC TRIALS AND TRIALS WITH 24 FUNDER INVOLVEMENT
25 Independent n=4 Funder involved n=76 26 Do not know Do not know 27 Yes No or Not Yes No or Not 28 Available Available 29 n % n % n % n % n % n % 30 31 Signed trial agreement with industry 2 50% 2 50% 0 0% 61 80% 14 18% 1 1% funder 32 33 Signed trial agreement included a 1 50% 1 50% 0 0% 45 74% 11 18% 5 8% 34 publication agreement* 35 Signed trial agreement included 0 0% 2 100% 0 0% 33 54% 19 31% 9 15% 36 presentation agreement* 37 Signed trial agreement included 0 0% 2 100% 0 0% 39 64% 17 28% 5 8% 38 confidentiality agreement* 39 40 Had access to the entire trial data set 3 75% 1 25% 0 0% 60 79% 8 11% 8 11% 41 42 Access used by those with access to 3 100% 0 0% 0 0% 53 88% 6 10% 1 2% 43 entire data** 44
45 1 25% 3 75% 0 0% 2 3% 67 88% 7 9% 46 Delay in publication due to funder 0 0% 4 100% 0 0% 9 12% 60 79% 7 9% 47 Disagreements with funder 48 Due to rounding the percentages may not add up to 100% 49 *n=2 and n=61 for independent and funder involved trials, respectively. Question was only available to those who answered yes to signing an agreement with industry funder 50 **n=3 and n=60 for independent and funder involved trials, respectively. Question was only available to 51 those who answered yes to data access. 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 42 of 42
1 2 3 4 5 6 7 APPENDIX 7. TABLE 3C 8 SURVEY REPORTED EXPERIENCE AMONG THE LEAD ACADEMIC AUTHORS STRATIFIED BY COLLABORATION STATUS 9 Collaborating with funder/ would in the Would not collaborate with funder again or 10 future does not know n=8* 11 Confidential:n=67 For Review Only 12 Do not Do not Yes No know or Not Yes No know or Not 13 Available Available 14 n % n % n % n % n % n % 15 16 Signed trial agreement with 56 84% 11 16% 0 0% 4 50% 4 50% 0 0% 17 industry funder 18 Signed trial agreement included 40 71% 12 21% 3 5% 3 75% 0 0% 1 25% 19 a publication agreement** 20 Signed trial agreement included 29 52% 20 36% 6 11% 3 75% 0 0% 1 25% 21 presentation agreement** 22 23 Signed trial agreement included 34 61% 18 32% 3 5% 2 50% 1 25% 1 25% confidentiality agreement** 24 25 Had access to the entire trial 57 85% 8 12% 2 3% 6 75% 1 13% 1 13% 26 data set 27 28 Access used by those with 49 86% 7 12% 1 2% 6 100% 0 0% 0 0% 29 access to entire data*** 30 Delay in publication due to 31 2 3% 64 96% 1 1% 1 13% 6 75% 1 13% funder 32 8 12% 58 87% 1 1% 1 13% 6 75% 1 13% 33 Disagreements with funder 34 *1 of the 8 authors would not collaborate with funder again the remaining 7 did not know. **n=56 and n=4 for collaborating with funder and would not collaborate with funder, respectively. Question 35 was only available to those who answered yes to signing an agreement with industry funder 36 ***n=56 and n=4 for collaborating with funder and would not collaborate with funder, respectively. Question 37 was only available to those who answered yes to data access 38 Due to rounding the percentages may not add up to 100%. 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj