Is this evidence of scientific fraud?

(the names of countries, journals, interventions and sample sizes have been changed to protect the potentially innocent and to avoid cognitive biases, invocations of  stereotypes and accusations that could lead to World War Z)

Two years ago the prestigious medical journal The Spleen published a randomized controlled trial that evaluated three therapies: genetically modified leeches (GML), standard leeches (SL) and mechanical blood-letting (MBL, standard of care) for acute complicated absentmindedness (ACA). This single center study randomized, in 1:1:1 ratio , 240 patients who presented at the Emergency Department of the Grand Fenwick Memorial Hospital and concluded that GML was associated with a 90% improvement in outcomes relative to SL and MBL. The lead author, a prominent absentmindedneologist and President of the corresponding Society of the Duchy of Grand Fenwick, concluded that GML should become the new standard of care and that SL and MCL should not be offered except as second line treatment for patients failing GML.

Upon closer inspection though, doubts emerged about the study: a meticulous reader noticed that there was an imbalance in the number of patients randomized: 105 had received GML, 70 SL and only 65 MBL. This was quickly pointed out in a heated letter to the Editor, to which the corresponding study author replied by dismissing this observation, attributing the imbalance to “randomization failure”, an unlucky coincidence  which had more patients were assigned to the GML arm. In the corridors of hospitals around the globe though, a more sinister explanation was offered: this in fact had been no randomized study of ACA in the Grand Fenwick Memorial Hospital but a simple observational study which the authors tried to “misbrand” as a randomized trial; after all Grand Fenwick research physicians had a reputation for publishing “positive” studies that only rarely could be replicated outside Grand Fenwick. Due to this uproar The Spleen issued a statement of concern and will likely be retracting the paper. While they are considering their decision, take a short poll to help them reach a decision

(the Bayesian answer will be given in a few days)

Edit
The expected number of patients in a 240 patient 3-arm trial is 80:80:80.
In effect we are asking whether this is so different from 105:70:65 as to constitute evidence for foul play.

Advertisements

One Response to “Is this evidence of scientific fraud?”

  1. Page Rev Bayes – we found statistical irregularities in a randomized controlled trial | Statistical Reflections of a Medical Doctor Says:

    […] Bayesian counterpart to the frequentist analysis of the Randomized Controlled Trial is in many aspects more straightforward than the Bayesian analysis. One starts with a prior […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: