Summary of SickKids work to date to address concerns regarding Dr. Koren's research
Concerns continue to be raised regarding Dr. Gideon Koren's published research and SickKids is committed to continuing to address these matters.
Toronto ON - Concerns continue to be raised regarding Dr. Gideon Koren's published research and SickKids is committed to continuing to address these matters. For a summary of the Hospital's current activities see: Update on Matters Related to Dr. Koren's Research.
1) Review of 1997 meta-analysis
Questions have been raised regarding a potential conflict of interest, based on the fact that pharmaceutical company Duchesnay Inc. Canada, the manufacturer of Diclectin - a drug approved by Health Canada to treat nausea and vomiting in pregnancy - is a source of funding for Motherisk programs and research, and some of the research and clinical work of Motherisk has focused on the use of Diclectin.
A controversy came to light when the media reported on a disagreement over the interpretation of one scientific paper published in 1997 by Dr. Koren and colleagues that reviewed and analyzed a series of papers studying the safety and efficacy of antihistamines in pregnancy to treat morning sickness. The disagreement arose when a 2014 publication could not replicate the findings of the 1997 publication.
In order to address the concerns, SickKids leadership asked SickKids' Research Integrity Advisor (who is responsible for reviewing any allegations of breaches of responsible conduct of research) to investigate the discrepancies. The investigation by the Research Integrity Advisor noted that there had been errors made in the citation of references in the 1997 meta-analysis and that the total number of subjects had been overstated, but found no evidence that data had been fabricated or falsified. The Research Integrity Advisor was able to confirm 174,568 subjects across 24 studies which corresponded with the totals derived from Table 3 in the 1997 meta-analysis. The Research Integrity Advisor did note, however, that the "over 200,000" subjects referenced in one part of the 1997 meta-analysis was an overstatement.
The review by the Research Integrity Advisor did not address the safety or protective effect of antihistamines during pregnancy. We believe this is a scientific debate based on different interpretations of data. To confirm this, SickKids asked an expert in the field to conduct an independent reanalysis of the original data in the 1997 meta-analysis.
SickKids retained Ottawa Methods Centre (OMC) of the Ottawa Hospital Research Institute to review the conclusions regarding the safety or protective effect of antihistamines during pregnancy. Whereas the 1997 meta-analysis suggested a protective effect of antihistamines against major malformations when taken for nausea and vomiting in pregnancy, the OMC review - based on the papers cited in the 1997 Koren publication - suggested that antihistamines were not statistically significantly associated with major malformations when taken for nausea and vomiting in pregnancy (i.e. antihistamines are neither protective nor harmful). At SickKids' request, Dr. Koren sent the OMC review to the journal. The journal declined to print the corrigendum given the length of time that had passed.
The authors of the 2014 study noted that they could only identify 22 of the 24 publications cited in the 1997 study and that they could only identify 139,414 subjects, rather than the larger number ("over 200,000" cited in one line of the paper; 174,568 subjects described in a table in the paper) identified in the Koren publication. They also questioned the validity of claims made in the paper regarding the safety and protective effect of antihistamines.
Since Dr. Koren had research funded by Duchesnay Inc. and has received payment from Duchesnay for various services, some media reports on this scientific disagreement suggested the possibility that scientific bias might have affected the outcome of the study.
2) Lang Review concerns with research papers referencing GC/MS testing
An additional research issue was identified during the Lang Review. Five published journal articles by Dr. Koren between 2000-2009 stated that samples analyzed with ELISA that had positive results were confirmed using GC/MS. However, SickKids learned through the course of the Lang Review that at the time these studies were conducted, Motherisk Drug Testing Lab (MDTL) did not send all positive samples to a reference laboratory for confirmation.
SickKids' Research Integrity Advisor reviewed these papers, but was not able to review the original data for the studies, as the data were no longer available. The document retention period had expired for four of the five studies. As per the SickKids policy for retention of data, data for the fifth study should have been available, but it was not. The Hospital's Research Integrity Adviser concluded that Dr. Koren failed to comply with the TriAgency Framework on the Responsible Conduct of Research in that he failed to: (a) keep complete and accurate records of data for seven years after the 2009 publication and; (b) use a high level of rigour in relation to these five studies. Subsequent to this review, the Hospital advised Dr. Koren, the Secretariat for Responsible Conduct of Research and the University of Toronto of the findings from the Research Integrity inquiry.
Dr. Koren has sent corrigenda to the journals stating that it was his understanding that some, but not all, positive samples were confirmed with GC/MS. Dr. Koren noted in his corrigenda that it was his opinion that this had no impact on the results and that conclusions remain the same. A corrigendum was published for three of the five articles (see links below).
Since SickKids does not have access to the original data, the statement that the fact that some of the positive samples were not GC/MS-confirmed had no impact on the results cannot be validated by the Hospital. Dr. Ronald Cohn, SickKids Paediatrician-In-Chief and Chair, Department of Paediatrics, University of Toronto, communicated this concern to the relevant journals. In respect of one of the articles, a Notice of Concern (link below) was posted to clarify that because the ELISA tests were only confirmed in about 10 per cent of the cases, there was still a possibility that the results obtained by ELISA might have differed from GC/MS results in some cases.
3) Review of active studies (post Koren retirement)
As part of the plan to transition active research studies upon Dr. Koren's retirement, the Research Ethics Board (REB) at SickKids reviewed all of Dr. Koren's open research studies. The majority of the studies were complete or inactive and were therefore closed. In the context of the review, the REB met individually with all collaborators and students working with Dr. Koren. In the course of those conversations information was uncovered and/or volunteered regarding noncompliance with REB standards. Several studies were then closed because they were not being done to the standards that we would expect; however, none of the issues identified would have had an impact on research subject safety or privacy.
The review found that there were two situations where records of calls to the Motherisk call centre were accessed by call centre research staff to determine eligibility for research studies without first going through the established process for research approval. The first project involved a student who searched the Motherisk database to obtain a list of all callers using Diphenhydramine, Dimenhydrinate and Cetirizin who contacted Motherisk between 2000-2014. This project was halted as soon as identified and no publication occurred. This study was not submitted to the REB, thus had no REB approval, and has no formal title. In the second case, a study entitled: Child's Developmental Milestones Following In Utero Exposure to Metformin had been submitted in July 2014 to the REB but data collection began prior to approval. No articles have been published based on either of these studies.
In total, we estimate that between 2000 and 2014, approximately 1,400 records of calls were accessed by call centre staff for research-related purposes, without first going through our established process for research approval. In the vast majority of cases, staff used the records only to determine whether a caller's information was relevant to research related to the use of a drug for diabetes, over-the-counter antihistamines or anti-nausea medications. In a small number of cases (approximately 70), having confirmed that the information was relevant, staff used the caller's contact information to seek consent for the use of the information and/or participation in a study and, where consent was provided, used the data for research purposes. In all of these cases, the use was without authority and contrary to SickKids' policies and privacy law. We verified that no caller information ever left SickKids and that the information was only accessed by staff working in the call centre. According to the SickKids REB, the board which reviews research applications, had the established process been followed, they would have likely approved the use of caller information with a waiver of consent. However, because this was not done the research was halted, and in the one case where a manuscript had been completed, it was pulled from a journal prior to publication.
Note, the privacy breach was reported to the privacy commissioner.
4) Conflicts of interest
Questions have been raised regarding a conflict of interest, based on the fact that pharmaceutical company Duchesnay Inc. Canada, the manufacturer of Diclectin - a drug approved by Health Canada to treat nausea and vomiting in pregnancy - is a source of funding for Motherisk programs and research and some of the research and clinical work of Motherisk has focused on the use of Diclectin.
While Director of Motherisk, Dr. Koren also consulted for Duchesnay. Motherisk has in the past provided information on Diclectin on its call lines and in its pamphlets, without consistently declaring Duchesnay's role as a funder of Motherisk.
Collaborations between industry and hospitals/researchers are critical for encouraging innovation that leads to the discovery of new health-care products/services/methods; however, such relationships may lead to perceived or actual conflicts of interest. Perceived or actual conflicts should be disclosed and managed. Through SickKids' annual conflict of interest declaration process, Dr. Koren disclosed funding received from various sources, including Duchesnay.
Dr. Koren's long-standing academic and funding relationship with Duchesnay was inconsistently disclosed on publications where it would be relevant, and his online profile noted that he consulted for Duchesnay. However, Duchesnay's support and the support of other companies was not consistently acknowledged on the Motherisk website and in several Motherisk pamphlets.
In a number of publications, Dr. Koren also acknowledged funding through "Research Leadership for Better Pharmacotherapy during Pregnancy and Lactation". This is a name Dr. Koren gave to an aggregate of donated funds in support of his research. These unrestricted funds were donated by a variety of individuals and organizations to be used at the discretion of Dr. Koren to fund his research. The primary donor of unrestricted funds in recent years was Duchesnay, pursuant to a contract between Duchesnay and The Hospital for Sick Children Foundation. In some cases where "The Research Leadership for Better Pharmacotherapy during Pregnancy and Lactation" is acknowledged, Duchesnay funding is also acknowledged. In other cases, it is not.
It has also come to the attention of the Hospital that in some articles published after Dr. Koren's departure from the Hospital, Dr. Koren has referenced an affiliation with SickKids.
SickKids has added disclosures to the Motherisk website and Motherisk publications, noting all corporate funding received by the program of over $5,000 between 2011-to 2018. On pamphlets/sites that specifically mention Diclectin, a disclosure statement noting Duchesnay's support for Motherisk has been added. SickKids asked Dr. Koren to stop using the "Research Leadership for Better Pharmacotherapy during Pregnancy and Lactation" designation for donated funds in December 2014.
We acknowledge that funding received by a hospital or scientist from industry may raise concerns for members of the public and the medical community regarding conflicts of interest and disclosures in publications should accurately and transparently reflect all relationships the author has with industry so that potential conflicts can be considered in interpreting findings.
Over the last three years, SickKids has updated its Relationship Management (Conflict of Interest) policy to ensure (a) it takes into consideration international health sector best practices as they relate to disclosure and management of conflicts of interest; and (b) that all relevant information is accessible and analyzed in a consistent manner across the organization. The policy includes a centralized approach to reviewing potential individual conflicts of interest. SickKids is now working with the University of Toronto and other Toronto Academic Health Science Network (TAHSN) hospitals to ensure a consistent approach to conflict management across the network. The SickKids policy is being used as a blueprint for relationship management across TAHSN. This work is significant and ongoing meaning that the TAHSN hospitals are working together to continue to advance our systems and approach to the management of relationships with third parties. Although we have traveled a considerable distance since this work began in 2015, we recognize that there is more work to be done.
The REB has also instituted a new process for more in-depth annual review of ongoing studies, as well as increased random and directed quality and privacy compliance audits.
The Hospital has also asked Dr. Koren to contact any journal where he has failed to acknowledge Duchesnay funding (where relevant) or published with reference to "The Research Leadership for Better Pharmacotherapy during Pregnancy and Lactation" to ask that the record be corrected. We have asked him to do this because journals are more likely to respond to a concern raised by an author than other interested parties.
In November 2018, SickKids asked Dr. Koren to contact each journal where he has published an article claiming an affiliation with SickKids since June 30, 2015 to correct the record and to refrain from making such references going forward. Letter to Dr. Koren from SickKids.