Scrutiny of publication bias increases
The Associated Press reports that court documents show GlaxoSmithKline (GSK) used a ghostwriting called CASPPER, which instructed salespeople to offer physicians help to write and publish articles.
Assistance ranged from “developing a topic” to “submitting the manuscript for publication”. According the Associated Press the memo said the aim of the programme was to “strengthen the product positioning and overcome competitive issues”.
Leemon McHenry, a ghostwriting expert, told the Associated Press that the unusual aspect of the GSK programme is the degree to which it is intertwined with internal sales and marketing.
The product referred to in the memo is Paxil (paroxetine), which is currently the focus of personal injury and wrongful death suits. GSK has said that any help was noted in the published articles, the programme was not heavily used and was discontinued a number of years ago.
Ghostwriting and bias in the pharma industry has become increasingly scrutinised, leading to Senators Charles Grassley and Herb Kohl proposing a bill that would make companies disclose payments of more than $100 (€70) to physicians.
Research into detecting bias
The release of details from the memo coincides with research published in the British Medical Journal into using statistical methods to detect bias. Researchers compared antidepressant data from the US Food and Drug Administration (FDA), considered unbiased, and journals.
Following the investigation the researchers believe that contour enhanced funnel plots could be used to identify and adjust for publication biases.
This bias could be a result of selective reporting, post hoc searches for statistical significance and applying particular analysis methods to the data.
The complete research paper can be found here.
