The end of evidence-based medicine?

Abstract

It is customary for physicians and surgeons to recommend treatments that have the highest probability of getting the patient a good outcome. Deciding treatment based on results (outcomes actually achieved) has been advocated for hundreds of years. We are now in an era when practice recommendations are supposed to be based on evidence from randomized controlled trials. There are many defects in relying on RCTs which are reviewed in this article. Advantages of observational studies are outlined, and the importance of assessing one's actual results when applying a treatment based on controlled trial evidence is set out.

Studies on the Infants Born to the Patients with Unexplained Recurrent Abortion Who Underwent Immunotherapy using their Husband’s Lymphocytes

“…The efficacy of the immunotherapy has been controversial [8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23], but a meta-analysis concerning the immunotherapy for patients with URA reported by the Recurrent Miscarriage Immunotherapy Trialists Group in 1994 [13], Clark in 2008 [7] and 2012 [14] concluded and Liu et al in 2016 [25] concluded that the immunotherapy may be highly effective. In addition, Pandey et al reported a significant improvement in pregnancy outcome in the patient group in their double-blind randomized trial of husband's lymphocyte immunization for women with recurrent spontaneous abortion [19].…”

“…Immunotherapy for these patients using their husbands' or a third party's leukocytes has been reported [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25]. Although the efficacy of this modality has been controversial, an analysis of controlled clinical trials revealed the efficacy for patients with Unexplained Recurrent Abortion (URA) [7,14,19,26,27], especially for those who were negative for blocking antibodies evaluated by a mixed lymphocyte culture reaction between spouses (MLR-BAbs) [19,22,[28][29][30][31][32][33]. The outcome of pregnancies, especially the condition of infants (such as the body weight of infants or the number of gestational weeks at which the infants were born) of patients after this immunotherapy, has not been fully elucidated, although some articles suggested absence of harmfulness of this immunotherapy for the infants born to immunized patients [34][35][36][37][38][39].…”

Popular myths in reproductive immunology

“…A funnel plot may reject the hypothesis of publication bias but lack the power to detect it (a type 2 statistical error). The second problem is that patients are rarely selected with sufficient rigor that one can include in a trial only those likely to respond to the test treatment (Clark, 2010(Clark, , 2012. A third problem is the absence of an independent test (e.g., improvement in a lab result that is related to the pathogenesis of the clinical disease) to show that the treatment was of adequate potency and corrected the putative pathogenic abnormality (Clark, 2009(Clark, , 2010(Clark, , 2012.…”

“…The second problem is that patients are rarely selected with sufficient rigor that one can include in a trial only those likely to respond to the test treatment (Clark, 2010(Clark, , 2012. A third problem is the absence of an independent test (e.g., improvement in a lab result that is related to the pathogenesis of the clinical disease) to show that the treatment was of adequate potency and corrected the putative pathogenic abnormality (Clark, 2009(Clark, , 2010(Clark, , 2012. In clinical trials of treatments to correct reproductive failure, the confounding effect of chromosomally abnormal embryos is rarely addressed adequately (Clark, 2003;Clark et al, 2004;Verlinsky and Kuliev, 2004;Munne et al, 2006;Fragouli et al, 2009;Santos et al, 2010;Liu et al, 2012;Bazrgar et al, 2013).…”

“…Fortunately, it has been shown that non-randomized cohort-controlled trials give the same result as RCTs (reviewed in Clark, 2010Clark, , 2012. In the ASRI meta-analysis of RCT and cohort-controlled trials of LIT in recurrent miscarriage, a maximum treatment effect occurred in double blind RCTs, the least effect in cohort-controlled studies, and an intermediate effect in non-blinded RCTs (Clark et al, 1997); meta-analysis of RCTs of LIT in RM has achieved sufficient statistical power to confirm efficacy (Clark, 2008(Clark, , 2012.…”