Georg Ivanovas From Autism to Humanism - systems theory in medicine
First of all there is the ‘human factor’, the obsequiousness of statistics to personal interests.
There is, of course, forgery (DiTrocchio 1993; Haffner/Lukacz 2003; Day 1998b). Nature magazine estimates that up to one third of the results are fiddled (Martinson et al 2005). Other estimates using a kind of peer judging suggest an even higher percentage (Fanelli 2009).
Another concern regarding the reliability of trails is the influence of the pharmaceutical industry (Mertens 2001; Moynihan 2003; Healy/ Cattell 2003). The the British Medical Journal published a series of articles demonstrating the manifold ways of how the pharmaceutical industry influences topics directly, or the medical journals publishing topics, or the journalists’ reporting, or the patients’ organisations recommending therapies (cf. BMJ 2003b). The former editor of the the British Medical Journal even called medical journals an “extension of the marketing arm of pharmaceutical companies” (Smith 2005b). For example, it came out that many articles ‘proving’ the benefit of hormone replacement therapy (HRT) had been written by ghostwriters of the pharmaceutical industy where the mentioned authors contributed little or no writing (Singer N 2009).
Industry sponsored results are much more likely to be positive. Negative results are mostly not published (Meier 2004). Rarely a medical journal attack results of the industry openly, as the Lancet did (Lancet editorial 2003). And rarely cases come into public how the industry tries to suppress undesirable outcomes (Baird/Thompson 2002, Savulescu 2004). Everybody is free to imagine how big the iceberg under such tips really is.There has been an attempt to improve the situation with an international committee registering all conducted clinical trails (De Angelis et al 2004). But after five years it still does not work (Mathieu et al. 2009).
Whether the situation will change through more law suites against pharmaceutical companies (Dyer 2004) cannot be predicted. The strategy of the pharmaceutical industry is to continue to advertise and sell drugs even when something has been scientifically proved to be harmful. And rarely political action is taken (Marmot 2004). Pfizer even continued to sell a drug, although already fined for illegal promotion (Lenzer 2004). Pharmaceutical companies bribe doctors (Harris 2004) and influence the medical market on all levels (Harder 2005). On conferences no speakers are admitted without the approval of the sponsoring drug companies. “Lectures are given on the same topic by the same speakers all over the world for years, and the free debate and exchange of contrasting opinions is, de facto, banned. It was not the same 20 years ago, but in the past 15 years this trend has become a rule, to the damage of medicine and freedom” (Vento 2002).
Another problem is the frequency of statistical mistakes in published studies. An analysis of papers published in Nature and the British Medical Journal showed that there were mistakes in up to a third of the papers, altering the significance in at least 12% of the papers (Garcia-Berthou/ Alcaraz 2004). As this concerns journals with a very high standard, where even statisticians control the papers to be published (Abbasi 2004), there seems to be no possibility to avoid this. It is to be expected that in journals with lower standards more faults occur.
In the cases where drugs have no major difference to the placebo, as, for example, in antidepressive drugs (Kirsch/Sapirstein 1998; Leuchter et al 2002) the results are manipulated with tricks (Pirisi 2002; Basskin 2003), or it is worked with percentages such that the results look impressive although the real effect is minor (Gigerenzer; Gigerenzer/Edwards 2003).
All this demonstrates that the presented knowledge is by no means objective or scientifically sound.