A few days after PositiveLite.com published the above-named article by T.J. Miller, a reader pointed out flaws in the author’s research and made it plain that we’d presented some bad information. PositiveLite.com has a strong commitment to accuracy and so we pulled the article right away, pending a review of the available data.
We apologize to our readers for our error but will add that we look forward to presenting T. J. Miller’s article once such a review has been made, taking into account his corrections (see below) as well as such new data as may have become available (also see below).
Rob Olver, Editor
On August 1, 2017 the author of this article and the editorial board of PositiveLite.com were presented with an objection to the content of my article, disputing the numbers that I reported for condom failure rates from the referenced source and also that I did not report in my article that the participants in the University of Alabama at Birmingham study (conducted by Macaluso, et al) self-reported an over-all “correct use, no problem” rate of 23.1% for male condoms, and 45.7% for female condoms. Also objected to was the reporting of data from the embargoed study, because it did not allow verification of the data contained within the study.
As a healthcare provider for 32 plus years, I understand the value of honesty, integrity and the absolute need for excellence in our work. What we as healthcare professionals do on a daily basis impacts people’s lives. So when these objections were raised I went back to my original work and re-evaluated everything. In doing this research I viewed the documents related to the 2007 American Journal of Epidemiology article (Volume 166 pages 88-96) that I had a hard copy of, to an online copy of the same article that I linked to within my original article. In reviewing the printed copy which I used to base my article on there are several errors as compared to the copy available online from the original journal article. I have made the determination after a line by line comparison that the printed copy that I used is corrupt and must be discarded.
So in using the data from the online magazine source, the following are the corrected results where a Prostate Specific Antigen (PSA) level was measured greater than 1 nanogram per milliliter for each condom type:
Male Condom: 86 times out of 635 total uses which equates to a failure rate of 14%
Female Condom: 100 times out of 599 total uses which equates to a failure rate of 17%
As one can plainly see these numbers are much lower that what I reported in my article. For this I apologize profusely and admit that a terrible mistake was made. I will have more about this mistake later in this correction.
The point being made by asking me to report the observational data from the study participants is quite simply that 23.1% of the participants using the male condoms reported no problem with use but only 14% had the PSA measurement above the cutoff (86% no problem rate). Similarly, those individuals who used the female condom reported no problems with use 45.7% of the time but only 17% had the PSA measurement above the cutoff (83% no problem rate).
To restate this in a less technical way, the observational reports of no problem with the device used was much lower than the actual measured failure of the device by detection of PSA so based upon this the devices performed better in reality than they did in the observations of the users. As I stated in my original article observational data is a problem with determining the accuracy/efficacy of condom usage, hence why the Macaluso study was undertaken. This data actually makes my point in certain ways by showing that in this particular study the participants thought that there were issues with the usage of the male or female condom when in fact the scientific measurements did not demonstrate a bodily fluid transfer by PSA measurement.
To respond to the objection raised about the use of data from an embargoed copy of research, this was a decision that I did not make lightly. It’s a valid point that because the content of the research was not disclosed, it did not afford the opportunity to view the research, how it was conducted, the methodology of the study or any other part of the study. My decision was made to publish the results based upon the following criteria. 1.) I am familiar with one of the researchers involved as I have heard this individual speak at educational presentations previously. 2.) I was assured by the lead researcher that the embargo would be lifted by mid-August 2017 and at that time I would be able to publish the full content of their publication and amend my article. 3.) I have found in several other cases where non-pharmaceutical or non-medical procedure health related articles were released with embargoed studies quoted, with the caveat that the embargoed information would be added at a later date, so long as it was supported by other studies with published source data.
Because of the fact that we now have a discordant set of results between the studies I am retracting my article fully until such time as the second study can be evaluated further. I think this is the fair thing to do for all parties involved. There is another reason for the retraction. I have become aware of an earlier article that my research did not find because it was only related to female condom usage, and my search parameters were set to male AND female condom, This article questions the use of PSA as a bio-marker for determining transmission of bodily fluids as it relates to a female only condom study published by Macaluso in 2003. The article I am referencing is https://academic.oup.com/aje/article-lookup/doi/10.1093/aje/kwf213 . Because of the discovery of this article I feel that a full retraction until a re-evaluation of my article can occur is warranted.
I would like to apologize to the readers of PositiveLite.com, the editorial board and the publisher for the error.