Guia GOLDACRE para manipular datos y lograr los resultados que quieres...

Choose an unrepresentative group
Goldacre says trials in one Cox-II inhibitor drug were conducted in an entirely unrepresentative group - people in their thirties and forties.
Younger patients are often more likely to respond positively to treatment, and less likely to show up any side effects such as cardiovascular problems.
Choose an ‘odd’ dose for your comparator drug
Specifically, use an abnormally high or low dose of the comparator to go against your drug ‘and hope that nobody notices’. Goldacre says there are numerous examples of this in studies of atypical antipsychotics.

Ignore side effects
Studies of SSRI antidepressants ignored the startling but clear sexual side effect seen in some patients taking the drugs - anorgasmia - the inability to have an orgasm. Goldacre says everyone surely agrees this is a serious side effect: “Wars have been fought, essentially, for the sensation of an orgasm.”

Peeking and stopping early
“It goes without saying that you have pre-specified critieria for stopping early and there are lots of situations where you would want to stop early,” says Goldacre.
“But that is very different to ‘peeking’ where you just have a look every now and then and if you see the results you want, you stop the trial and write it up in a hurry.”
He adds: “You could make a bit of a case that this happened with the [Avandia] RECORD trial - that is certainly Steven Nissen’s argument - and not just his actually. There is a huge debate around whether the RECORD trial broke its stopping rules and was written up inappropriately, and, in response to work that was casting doubt on whether rosiglitazone was a good idea.”
Reference: J Am Coll Cardiol. 2010 Feb 2;55(5):428-31.
Randomized trials, statistics, and clinical inference.
Stone GW, Pocock SJ.

Ignore drop outs
These trial participants are more likely to have done badly in the trial, and more likely to have had side effects. Goldacre says trial sponsors still ignore them, and do not follow up on these cases or include them in their final results.

Clean up the data
Pruning unfavourable ‘outliers’ in the data will help clear up ambiguity in the data, so the temptation to exclude them is strong.

Torture the data
Sub-group analyses will eventually yield positive results.

Bury bad data
A 2008 paper published in the NEJM uncovered large amounts of negative data on SSRIs which had gone unpublished, altering the perception of the drug.
Selective Publication of Antidepressant Trials and Its Influence on Apparent Efficacy
www.nejm.org/doi/pdf/10.1056/NEJMsa065779
Produce duplicates of positive trials
Tramer (BMJ 1997; 315 : 635).

Ben Goldacre studied medicine at Magdalen College, Cambridge and is currently a research fellow at the London School of Hygiene and Tropical Medicine.
He has written the weekly Bad Science column in The Guardian since 2003, tackling subjects as diverse as the Catholic church’s opposition to condoms, the MMR scandal, self-appointed nutrionist Gillian McKeith, and of course, dubious practices in the pharmaceutical industry.
His Bad Science book (published by 4th Estate) has sold 240,000 copies, reached No.1 in the paperback non-fiction charts, and has been published in 18 countries.
He is currently working on a follow-up to the Bad Science book, and is interested in taking part in a lively public debate with the pharmaceutical industry.

He can be contacted at: ben@badscience.net