But closer look reveals there are significant problems with the reliability and completeness of the data
By Randall K. O’Bannon, Ph.D. NRL Director of Education & Research
Editor’s note. On Monday Dr. O’Bannon outlined in detail the case made by two pro-abortion researchers that women are safer if they have their abortion via webcam than if they actually come to an abortion clinic where they will see an abortionist. He outlined some of the many shortcomings of “Safety of Medical Abortion Provided Through Telemedicine Compared With in Person” and will conclude today with still further flaws in the study by Daniel Grossman and Kate Grindlay of the University of California, San Francisco (UCSF).
Consider the Source
Randall K. O’Bannon, Ph.D.
The data that Grossman and Grindlay rely upon comes from reports made by Planned Parenthood and information they relayed to Danco, distributors of the abortion pill. Up until March of 2016, the U.S. Food and Drug Administration (FDA) required that prescribers report all transfusions, hospitalizations and serious adverse events to Danco, who were to relay those reports to the FDA [1] . But according to Grossman and Grindlay, now, under the new protocol, only deaths need be reported; the complications recorded here would have been under the earlier, broader FDA standard.
The researchers here take for granted that Planned Parenthood employees will be forthcoming with details about complications. They write that “Planned Parenthood staff are trained in accurate and complete reporting of medical abortion-related events” and give assurances that these reports are audited by, well, the Planned Parenthood Federation of America.
This might be impressive for people conditioned to support the word of Planned Parenthood. But it far less straightforward for those who are familiar with the history of Planned Parenthood, and PPH in particular, with regard to the abortion pill.
For example, Planned Parenthood’s Iowa affiliate was part of the original U.S. trials of RU-486. They notoriously declared in March 1995 that the trials had ended with “no complications” among the 238 women who had been part of the trials there.
But a doctor who had saved the life of a woman who was in the trials and who nearly bled to death came forward to tell the Waterloo Courier (9/24/95), “If near death due to the loss of half of one’s blood volume, surgery, and a transfusion of four units of blood do not qualify as a complication, I don’t know what does.”
Missing Data from the ER
Even if Planned Parenthood intended to be open and above board, there is no guarantee that they would be made aware of every “adverse event.” A woman who simply goes to the ER, or her regular doctor, and does not inform the clinic or return phone calls, will not be recorded in these reports.
This is going to be a special problem when, as in the case of many of PPH’s webcam locations, these are not regular clinics with full medical staff or equipment and when their instructions to women were to seek help at their local ER in circumstances of significant bleeding, etc.
This is a significant enough issue that even Grossman and Grindlay sought to take steps to address it. They say they attempted to contact all 119 emergency departments in the state to ask if they had had any women show up with possible complications from a chemical abortion.
There are several problems with this research strategy. Some the UCSF team recognize, some they don’t.
Only 35% of the emergency departments responded. Most said they hadn’t encountered any such patients, but some said they simply weren’t sure.
However 71 of those ERs (or nearly 60%), didn’t reply. Even if one assumed those that did respond were representative, there is a lot of missing data here.
Almost two-thirds of your data set is missing–that, in and of itself, is enough to raise the concern that the published results here may not be reliable.
Furthermore, without more on the geographical spread of the responding ERs versus the non-responders, there is no guarantee that any or even some of the responding ERs were in the rural areas where the webcam abortion program was set up. Those are the locations where the webcam patients with problems would have shown up.
If the handful of ERs that responded to the survey were largely from metropolitan areas where PPH had full abortion clinics performing in-person chemical abortions B clinics that could have, or should have been equipped and staffed to deal with the complications themselves B there would likely be few if any cases showing up at those area ERs. Any reports would have reflected that.
The Problem of Unshared Information
It gets worse.
We don’t know specifically that it was the case with women aborting there in Iowa with Planned Parenthood. However those who have seen how the promoters of the abortion pill operate know that some groups have explicitly told women they don’t need to tell the doctor at the ER that they’ve taken the abortion pills and are having a reaction.
They are told that they can simply say they are having a miscarriage and that there won’t be any way for the doctor to tell any different.
Whether PPH told women this or not, it is not difficult to imagine that a woman might not tell her ER doctor she had taken the abortion pills. If that is the case, the ER doctor would not otherwise know and would not report it as an abortion complication.
A survey asking that ER doctor if he or she had encountered any woman in the past 12 months “with a possible complication of medical abortion” is not going to turn up any useful data. Given this very real possibility, it would be totally illegitimate to present their lack of knowledge of any such cases as proof that there had not been any.
Years of Missing Data
The UCSF team’s study is supposed to have covered PPH chemical abortion patients from July of 2008 to June of 2015. But the ER survey they performed from June to October of 2014 inquired only about the most recent 12 months period.
The authors say that they only asked ER doctors about cases they remembered from the previous year to avoid possible recall bias; the authors grant doctors may have already been relying on memory rather than actual hospital records.
However it is not specified how they used this information to develop data for the five years of webcam abortions that occurred before the single year about which UCSF asked the ER doctors. Also unclear is how they used that same 2014 data predict actual future adverse events for the eight months of 2014 and 2015 that followed the end of the ER survey!
In any case, given inherent problems with the data and data gathering, it strains credulity to claim definitive low complications rates for webcam abortions based on reports from the clinics that sold those chemical abortions and a handful of ER surveys that may not even be from emergency departments who would have seen or recognized chemical abortion patients with problems.
Grossman and Grindlay admit there are problems with the ER data. They write, “The survey of emergency departments is significantly limited and results must be viewed cautiously.” Nonetheless they still claim that their study proves that concerns about the safety of webcam abortions are unfounded.
Dismissing Their Own Doctors
One of the biggest problems is that Grossman and Grindlay end up arguing that webcam abortions are safer than those mediated by a doctor. If they wish to defend their data, what this says about the medical service provided by abortionists is astonishing. That is, that a woman seeing them in person for her abortion pills, undergoing a physical exam there at the clinic, having a full face to face consultation, is more likely to have a complication than a woman who only converses with the abortionist over a computer monitor.
This either reveals something ominous about abortionists and their clinics, or it is an indication that something is off with the data. Neither would offer the sort of validation or support that the abortion industry seeks from this study.
Missing Demographics
Finally, there is one other problem with the UCSF study. It lacks any demographic or sample data.
What was the racial or ethnic mix of the patients in the study? Was there any demographic difference between those who used the webcam rather than the “in person” chemical method?
There may be racial or ethnic differences between urban and rural populations (webcam abortions designed to cater to those farther out from the metropolitan areas where the full service clinics are found). There could also conceivably be racial or ethnic factors in there being more complications with one method than another. That Iowa has a less diverse population than, say, California, might itself lead to different results.
The same can be said for age and education. Are the “superior” webcam numbers due to the women in rural areas being older, or having more education? Hard to tell without data, even if these are critical factors.
How far along they are with their pregnancy is also a factor. It is widely recognized that the “success” of the chemical method declines each week the baby develops, and it is also generally believed that the bigger the baby, the more likely the complications, so any differences between the webcam group and the “in person” group with regard to gestation will also be statistically relevant.
What We Do Know
We know that not every woman who undergoes a chemical abortion dies. But it is no less true that each woman who aborts faces painful cramps, heavy bleeding that goes on for some time and exposes them to serious risks. That’s bad enough, and warrants some degree of responsible medical supervision.
Some, failing to get knowledgeable and timely help, have indeed died, bleeding to death, suffering from a ruptured ectopic pregnancy, or contracting a rare infection that thrives in hidden open wounds.
In the end, the study offers a tacit acknowledgment of some of the risks, but its methodology is too flawed to full accounting of the complications specifically connected to webcam abortions or to chemical abortions generally.
[1] In March of 2016, the FDA eliminated a follow up visit for the prostaglandin misoprostol, which may now be taken at home. The FDA also reduced the required dose and loosened the prescriber requirements to a wider range of medical personnel.