The e-Journal of Neonatology Research received the following letter this morning:
Dear Dr. Gordon,
I read with concern your recent article posted on the website of the e-journal of Neonatology Research (now retracted).
I am the lead author of one of the studies you reference in your article (Freed GL, Dunham KM, Clark SJ, Davis MM. Research Advisory Committee of the American Board of Pediatrics. Perspectives and preferences among the general public regarding physician selection and board certification. J Pediatr. 2010 May;156(5):841-5). My concern arises as several of the statements you make regarding our peer-reviewed manuscript are factually in error. These range from basic misinformation such as the fact that you have cited the paper incorrectly; it was not published in a supplement to Pediatrics as you report, but actually as an original article in a regular edition of Journal of Pediatrics (J Pediatr. 2010 May;156(5):841-5). Although normally this in and of itself would not be a significant concern, you actually make a (false) point in your article that our study was “written in the format of a research article and published in Pediatrics within a supplemental issue (i.e. a paid for publication).” As noted, this is a false statement and a false assertion; the paper was not published in a supplement, nor was it even published in the Pediatrics. Your statement is clearly an attempt to discredit our work. However, it is ironic you have made fundamental errors if your effort to discredit us.
There are several other factual errors in your manuscript. Most of the correct information is clearly stated in the manuscript we published. Based on your incorrect attribution of the study noted above and the points of error noted below, I am doubtful as to whether you have actually read the paper. Regardless, I have listed below your additional errors with specific corrections for each:
1. “The principle basis for this argument was a commercial survey”
This was not a commercial survey. This was an academic survey conducted by an established and experienced research group led by Drs. Gary Freed and Matthew Davis at the University of Michigan. Dr. Freed was the principal investigator of the project and Dr. Davis leads the team which conducts the C.S. Mott Children’s Hospital National Poll on Children’s Health (NPCH). The Poll is widely recognized for its rigor, relevance, and impartiality. Many peer reviewed publications have resulted from work conducted by the NPCH (http://www.med.umich.edu/mott/npch/index.htm or www.chear.org). The content and design of all NPCH surveys are totally and completely under the control of the investigators (NOT Knowledge Network). The NPCH is fielded through the KnowledgePanel, maintained by the widely respected survey vendor Knowledge Network.
2. This was a “paid for study for the ABP”:
The study was funded by the ABP Foundation as an extramural grant to the University of Michigan. The ABP did not have editorial control over the survey, study findings or presentation of results. We are an academic research unit (The Child Health Evaluation and Research Unit; www.chear.org) at the University of Michigan with a strong national and international reputation for rigorous and independent health services research. Your statement suggests the ABP funded the study and directed the methods and results. This was not the case. We receive funding from many sources including the NIH, CDC, WHO and multiple other federal agencies. We do not accept funding from for-profit companies (e.g., Pharma). Although public agencies and foundations fund our projects, they do not direct our work. The same is true for the ABP.
3. The participants were “disproportionately educated and financially well off parents (i.e. a sample cohort that did not well represent the general population).”
Again, this is simply false. Sampling for all studies conducted by the NPCH (including this one) is strongly representative of the national demography of parents and is neither disproportionately educated (43% with education of high school or less) or financially well off (24% with annual income <$30,000). Data regarding the sample for this study can be found in Table 1 of the article. Again, it appears you were not looking at our article when writing your critique of our work.
4. “We note that Knowledge Networks has suggestive statements on their website that are consistent with this: “Knowledge Networks helps … clients build brands and drive business by leveraging sophisticated analytic techniques” and“…insist that the management destination drives the segmentation design and analytics.”
While Knowledge Networks(KN) is an established vendor for marketing surveys, that side of their business is maintained separate from their academic survey work (like NPCH). Furthermore, KN does not have any editorial input for the design and content of any NPCH surveys; they are solely the survey vendor (i.e., they simply distribute the survey).
5. “ABP did not publish this survey as a brand analysis; it was written in the form of an original research article published in Pediatrics as a supplemental issue (i.e. a paid for publication).”
Again, the article was published as an original research article because it was original research that passed muster in scientific peer review. It was published in J Pediatrics, not in Pediatrics in a supplement as you falsely claim.
6. “The one unloaded question that they printed in the publication was the only one that actually addressed whether or not failure to complete MOC might affect the parents’ decision to bring their child to their current care giver (answers were “very likely, likely, not likely, not at all likely”). The trends were not as robust in this answer format when compared to those questions that were loaded, but there was a trend towards respondents’ desire for MOC.”
There is no measure of “robust” that we present. That is actually not an appropriate use of the term in this context. Rather, we let the data speak for themselves. In this case, it appears more than a “trend”: Table 4 of the manuscript shows that 77% of parents say they are likely to change doctors if their children’s physicians do not keep up with certification.
7. “There was no specific information on the survey describing what MOC consisted of (and the survey was done before MOC existed). Based on this one questionable piece of “research”, the ABP has repeatedly claimed that the public demands MOC in public forums.”
We certainly take exception to your comment that we conduct “questionable research”. We believe that such a comment is uncalled for and unprofessional. Further, at the time the survey was fielded, there could not have been specific information provided about MOC because it had not yet been codified. Rather, we specifically included measures in the study that identified what we believed to be potential components of MOC (eg, completing written test at regular intervals, checking for quality of care, ratings from patients and parents), all of which were supported by the public in this study. These components are now actually a part of the MOC process.
In addition, there is one minor point that deserves clarification:
8. “The arrangement of the answers also had the potential to create survey bias because the number of positive options outweighed the options with negative connotations when referencing the topic of interest (“unsure” is a neutral answer)”
While this critique is theoretically grounded, our pilot-testing and full field-testing indicated that it was appropriate to gauge public opinion in more detail on the positive end (eg: important and very important) and that more detail on the negative end was superfluous.
We believe the false statements regarding our work you have published constitute a significant and serious breach of professional ethics. However strong your feeling may be on a particular issue, it is inappropriate to make false statements in the medical literature and to falsely malign others in the academic community. As such, we respectfully request that you retract your article and publically acknowledge the false statements and false aspersions on the credibility of our research team (the Child Health Evaluation and Research Unit at the University of Michigan and the C.S. Mott Children’s Hospital National Poll on Children’s Health) you have made.
Sincerely,Gary L. Freed, MD, MPH The Percy and Mary Murphy Professor of Pediatrics Director, Division of General Pediatrics Director, Child Health Evaluation and Research (CHEAR) Unit University of Michigan Health System 300 NIB, Room 6E10 Ann Arbor, MI 48109-5456 Phone: 734-615-0616 Fax: 734-764-2599 Email: firstname.lastname@example.org www.chear.org
As the author and Editor-in-Chief,
I read this email with a pit in my stomach. It is a fine line when writing an introduction, one tries to write words that compel the reader to a point of view but at the same time remains squarely within the factual realm. This was a gut check. As I went back and researched the manuscript one central fact became clear. I had a typographical error in the reference list. Dr Freed’s study had in fact been mis-referenced. I had inadvertently fused the original reference with a subsequent ABP reference (which was from a different journal and was a supplemental). In rewrites, this typo became a fact and I worked it in as yet another reason I didn’t trust this study. But I was wrong on this and several other points. It was a peer-reviewed study. It was not commercially designed. It was demographically appropriate. To make matters worse, although the manuscript has still not been published, the address from a very early draft was leaked. I had edited certain items within the introduction long before Dr Freed contacted me, (although faults still remained). Some of these original, already corrected items were uncorrected in the version he was quoting to me. This means that a very early version of this manuscript has become widespread in the electronic literature.
I have communicated with Dr Freed and am in agreement that the only way to be fair and to help repair the reputation of his research unit is to withdraw the paper (even before it is published), redact the errors, then resubmit the study as a new manuscript. I will also list here the inaccuracies that earlier drafts of the manuscript contained regarding his study:
1. The manuscript in question was not a supplemental publication in Pediatrics, but an original research study published in the Journal of Pediatrics.
2. The study was not designed by a private company (only administered by it).
3. The study was not supported by the ABP (rather it was supported by the ABP foundation).
4. The study participants were not socio-economically skewed.
I apologize to Dr Freed and his group for these unfortunate events and hope that those sites currently carrying the old version of the manuscript will publish this retraction, as well our newly submitted version (which can be found here: http://www.neonatologyresearch.com/?page_id=1347).
We are also going to discuss policy changes with regard to prepublication manuscripts at eJNR.
Sincerely,Phillip Gordon Author Editor-in-Chief