When EBM is not really EBM

This post might seem a bit off-topic, but it's relevant to anyone who is interested in any type of human research. As you might be aware of, I'm a strong believer in evidence based medicine (EBM) and am currently studying for a Masters in Clinical Epidemiology. I find it scary that there even exist doctors who are against the concept of EBM (see http://www.kevinmd.com/blog/2011/10/evidence-based-medicine-removes-physicians-autonomy.html). However, I've also become aware of the highjacking of EBM to suit researchers' or governments' agendas. I have sometimes been stuck by the contradiction between what I am being taught and what is being touted as 'accepted evidence' or a 'valid' study the next minute. Of course researchers are human and we all have our opinions and beliefs (not to mention prejudices) that will spill over into research. Many researchers become interested in a topic because of a personal experience which sparked an interest in contributing to or improving research in that area because they had gained a personal insight into the problem that previous researchers missed. Improving patient care is a great reason to support human research and EBM. On the other hand, there are researchers who, motivated by personal gain whether this is career advancement, religious/'moral' beliefs, or financial gain, manipulate research using their credentials as 'evidence based researchers' to bias research outcomes. This is the complete opposite of what they are claiming to be doing, to the great detriment of EBM. Sometimes fear underlies their motivations because their research or investigation could show something the government or their peers did not want to acknowledge, resulting in law suits, medical scandals, or ridicule and it could spell the end of their career.  The following article from the Indian Journal of Medical Ethics outlines some of the ways research can be exploited to 'prove' what the researchers want it to prove:

Evidence based medicine: can the evidence be trusted? by Prathap Tharyan
Indian J Med Ethics.2011 Oct-Dec;8(4)

Link to full text article:

http://www.issuesinmedicalethics.org/194ed203.html

No one can claim to have no agendas whatsoever when it comes to conducting their particular research. One could argue that researchers should be able to design and analyse their own research according to what they have been taught about EBM to reduce bias and to critically appraise research undertaken by others by those standards. I'm not sure all researchers are able to do this, particularly if they have been taught about scientific thinking after their formal education, not as part of it. It takes a special mind to question what professors may be teaching you are established 'facts' and no one wants to be accused of 'arrogance'. However, not everyone is capable of scientific thinking, lateral thinking or thinking outside the box, no matter how successful, experienced, skilled or well-respected they may be.

For some areas of Gynecology and Obstetrics (such as Asherman's syndrome) it has not even reached the stage where EBM is being used much at all, so its manipulation is not the biggest problem. The first hurdle is to actually apply EBM to the field, which is desperately needed. There are however some areas where I can see some manipulation (not necessarily machiavellian) already, for example studies about misoprostol where low doses and short periods for assessing outcome are being used to 'prove' how poorly it performs compared to D&C. Another pitfall I see is something called 'intention to treat' or ITT. Although ITT is a non-controversial and accepted concept in EBM and I agree with all other aspects of EBM, this is one area that I question. ITT means that in a randomized controlled trial (RCT) those assigned to the experimental group (new treatment being tested) should be analyzed according to belonging to that group, and those assigned to the control (no treatment or standard treatment) should be analyzed according to the control group, REGARDLESS of whether they actually undergo those treatments. The rationale behind this is that, if for example 80% of the people in the experimental group decide not to take the new medicine, the drug will not be a success no matter how well it works because of patient non-compliance. This could be due to severe side-effects or inconvenience (you have to inject it whereas the standard drug can be taken orally). While I can understand its usefulness in assessing feasability of a treatment, it can be very misleading about actual outcomes, efficacy and even causal mechanisms. Imagine a study comparing the outcome of intrauterine adhesions (IUA) following miscarriage management by misoprostol or surgical curettage (D&C). If a low ineffective dose of misoprostol is dictated by the study design, or if the researchers decide that treatment will be evaluated after 48 hours instead of a week or two, or if they mistakenly diagnose retained products of conception by ultrasound (a common occurrence) after misoprostol treatment in a large portion of participants, these women will automatically be given the standard treatment which is D&C- in this case, the exact same treatment as the 'control' group! This means that study results according to ITT could inaccurately find that misoprostol 'causes' IUA and infertility, rather than that the misoprostol regimen was not given a fair go of success. In fact, even with irreproachable statistical methods, ITT can find that there is either 'no significant difference' or a significantly 'higher' rate of IUA among women in the misoprostol treatment group than in the D&C group depending on chance or on inclusion criteria (they may have included women with past D&Cs in both groups. See the MIST study). Although the logic is flawed, researchers could probably get away with arguing that this was 'proof' that IUA was not necessarily caused by instrumental injury to the endometrium and draw other naive conclusions, preventing progress in the field or prevention of the condition.

The point of this post is to highlight that there is yet a long way to go in achieving trustworthy EBM, but it should remain the goal of all medical research and practice. Another important thing to note is that what may be accepted as medically 'sound' today (due to poorly designed studies masquerading as evidence based research) could appear ridiculous in a century- and vice versa. A good example of this is the Hungarian Obstetrician Ignaz Semmelweis, who first understood that the cause of 'childbed fever' (puerperal fever) was contamination by Obstetricians with unclean hands. How did this observation he proved affect his career? He was shunned by his peers who were 'offended' by the suggestion they wash their hands ("the nerve of him!"), and his views were dismissed ironically as being 'unscientific' (despite his experimental evidence that washing hands reduced mortality rates to below 1%) only because he was unable to explain why. It was only after his death that knowledge about microorganisms explained the association (the germ theory of disease), and his ideas about hand hygiene became understood and accepted into practice. He is one of the doctors whose rationale led to the emergence EBM, but he died alone  (ironically of septicemia) with no glory, in a mental asylum (because of course anyone who rallied against accepted dogma was 'mentallly ill', see Foucault's biopower theory) and betrayed by all. I can see why some researchers would rather go with the flow, but to me Semmelweis was a courageous hero, ahead of his time in his thinking, while his detractors thought they were being scientific by parroting ideas they didn't fully comprehend ...or simply were not brave or caring enough about their patients.

Insights from an Asherman's syndrome expert (Part II)

Apologies for the long delay... To recap, in my last blog post I gave a brief background about Dr Charles March of California Fertility Partners and his vast experience on Asherman’s syndrome (AS), which he summarizes in his last review, Management of Asherman’s syndrome (1). Dr March’s article emphasizes many important issues that are often overlooked and reconsiders the validity of accepted beliefs. It’s clear throughout that he has given a lot of critical as well as practical thought to different aspects of Asherman’s syndrome.  The review starts off with a background including the epidemiology of AS, symptoms and diagnosis, then management, which he summarizes by the acronym PRACTICE, for prevention, anticipation, comprehensive therapy, timely surveillance of subsequent pregnancies, investigation (potential treatments like anti-adhesive gels and vitamin supplementation to improve blood flow), and continuing education (continuing medical education courses and literature).

I’ll focus on the parts of the review containing relevant insights which I have not encountered in other reviews, or which are important enough that repeating them is warranted.

General

1.       Asherman’s syndrome is not rare. He points out that most physicians believe that AS occurs  rarely and fail to diagnose it even when a patients exhibits obvious symptoms. Table 1. (page 64 of the article) displays convincingly that AS is not rare by showing the prevalence of AS in different populations of women undergoing hysteroscopy, and the incidence of AS after various procedures

2.       The term ‘syndrome’ does not correctly describe the condition. This is because the condition has varying symptoms and pathology, from endometrial sclerosis without intrauterine adhesions (IUA), to IUA without endometrial sclerosis. However, he points out that Asherman’s syndrome encompasses all of the different possible manifestations of endometrial injury, from all possible causes and varying symptoms. For this reason he continues using the term instead of IUA or traumatic amenorrhea.

Etiology

3.       AS most commonly occurs after dilation and curettage during or shortly after a pregnancy. However, diagnostic curettage may also lead to AS. Ignorance about this continues. I’ve heard several doctors who are not experts on AS and probably haven’t read the literature on it claim that the condition only occurs where there has been pregnancy and never in its absence. This old belief has been proven to be untrue. Interestingly, Dr March mentions that curettage used to be used during laparoscopy to investigate infertility, observing that this probably caused more harm than good. It is very refreshing to hear doctors reflect on their predecessors’ or colleagues’ mistakes and admitting that sometimes what seems like a progressive approach turns out to be a step backwards.

4.       There is no evidence that uterine malformations such as Mullerian anomalies are more prone to developing adhesions, even though there is very strong evidence that the two are associated. He points out that the correlation can be explained by the high miscarriage rate- and subsequent D&Cs- among this group of women. Another reminder that correlation is not evidence of causation. The bottom line is that injury from surgical trauma (or TB in some countries) is what causes AS, and there is a tendency for some authors to mystify the condition by neglecting the obvious associated risk factor i.e. women with Mullerian anomalies tend to have more D&Cs which lead to AS.

5.       There is no evidence that infection leads to AS. Dr March notes that infections occurred in less than 1% of his patients. The Polishuk case series (2) showed that among 171 women who underwent C-sections, of the 28 who had severe endometritis, none developed de-novo adhesions (the one patient who did develop cervical adhesions already had a previous history of AS from several miscarriages treated by D&C (her AS was also ‘treated’ by D&C).  Furthermore, there are no studies whatsoever that show any evidence that endometritis alone leads to AS.

6.       Endometriosis may develop in AS with outflow obstruction and patent fallopian tube(s) if treatment is delayed. This is an important point that isn’t mentioned enough in my opinion. It’s also another good reason why AS should be prevented or if it's too late, treated to prevent further problems: it can lead to 2 fertility problems for the price of one (a bargain nobody wants).

7.       Inexperienced operators can inadvertently cause uterine damage. New instrumentation and technologies for intrauterine surgeries are proven to be safe by experts, however they may not be so harmless if inexperienced operators use them.

Diagnosis

8.       Menstruation and withdrawal bleeding cannot rule out AS. Some women with IUA have normal periods (‘eumenorrhea’) and some women with amenorrhea have withdrawal bleeding after hormone administration. Therefore hysteroscopy is the gold standard for diagnosis. (HSG and SIS can lead to false positives). However, the opposite is true: women with hypomenorrhea or amenorrhea after intrauterine surgery are likely to have IUA.

9.       HSG and SIS can rule out IUA, but cannot rule out endometrial sclerosis, which can be diagnosed by hysteroscopy. On the other hand, one important advantage of HSG is that it provides information on the patency of fallopian tubes, which hysteroscopy does not. Single or bilateral tubal obstruction is in some cases caused by AS if adhesions occlude the ostia/ostium, so it is important to confirm after corrective surgery that the adhesions have not reformed. A partial or complete obstruction of one or both ostium can lead to tubal pregnancy or infertility.

Next time: Dr March’s advice about prevention, treatment and monitoring of post-AS pregnancies.

References

March C. Management of Asherman’s syndrome Reprod Biomed Online. 2011 Jul;23(1):63-76.

Polishuk WZ, Anteby SO, Weinstein D. Puerperal endometritis and intrauterine adhesions. Int Surg. 1975 Aug;60(8):418-20.