DEBBIE ELLIOT, host:
Here in the U.S., most pharmaceutical research is funded by the drug companies who stand to profit from the results. The companies frequently hire universities to study whether a drug is truly safe and effective, because universities offer both expertise and a degree of impartiality. But drug companies often exert a great deal of control over how university findings are interpreted and presented. A British scientist has been visiting Washington to air a dispute he's having with a big U.S. Company. NPR's Richard Harris has the story.
RICHARD HARRIS reporting:
This story involves a drug that's approved to treat bone thinning, osteoporosis. Dr. Aubrey Blumsohn and a more senior colleague at the University of Sheffield in England got paid by the manufacturer, Procter & Gamble, to perform more detailed studies about he drug Actonel. Procter & Gamble was looking for information it could use in marketing its drug against it's main competitor, Phosomax. Blumsohn's job was to study thousands of blood samples taken in a previous study which compared Actonel to a dummy pill.
Dr. AUBREY BLUMSOHN (Sheffield University senior faculty member): We didn't know which of the samples we were measuring were derived from patients getting the drug Actonel, and which patients were getting placebo and we also didn't know which patients during the study had vertible fractures, in other words, fractures of the spine.
HARRIS: And so it should be. Scientists doing lab studies shouldn't have any information that might bias their conclusions. Blumsohn sent his lab results off to Procter & Gamble. The company did all the fancy statistical analysis of the results and sent back a detailed summary of the findings. Blumsohn was not satisfied with just a summary, when his request to be sent a complete set of all the company's raw data was rebuffed, according to emails Blumsohn gave NPR.
Dr. BLUMSOHN: As an academic author, you vouch for the veracity of the material written in your name and this was obviously impossible under the circumstances, and I became really quite frustrated. And eventually after having raised the issue of access to data several times the company invited me down to a meeting with one of the company's statisticians at the U.K. headquarters of Proctor and Gamble.
HARRIS: Blumsohn spent a whole day reviewing the data and the methods Proctor and Gamble was using. He was startled to learn that the analysist ended up throwing away about 40 percent of the data.
Mr. BLUMSOHN: That 40 percent of the data was a critical part of the data in terms of the competition between these two particular drugs.
HARRIS: Why didn't you just say take my name off of this, this is no longer my work.
MR. BLUMSOHN: I'm not convinced that is an appropriate route to take. That scientists who feel uneasy about a particular type of research and feels that a particular type of research may be forgered into inappropriate shouldn't perhaps withdraw it from the process and allow others to simply do the business. So I didn't feel that was an appropriate thing to do.
HARRIS: So instead he has mounted a public campaign against his senior collaborator at the University of Sheffield, the university itself, Proctor and Gamble and a medical journal that published a previous study on the drug using the analysis method he disputes. Proctor and Gamble says it is a regrettable misunderstanding over arcane statistical methods. Company spokesman Tom Millikin says the company has now agreed to let a statistician chosen by the University of Sheffield review all the data and methods.
TOM MILLIKIN (Spokesperson, Proctor and Gamble): We stand by the research and are confident that the outcome will confirm that our statistical methodologies and researching conclusions were sound.
HARRIS: Conflicts over data access are actually fairly common. Dennis Black, an epidemiologist at the University of California San Francisco says that's partly the result of the way academics and industry have come to collaborate.
DENNIS BLACK (Epidemiologist, University of California): The common practice that researchers are writing papers and putting their names as author on paper without having access to the raw data, they often will see the analysis that have been done by someone else but it's not uncommon that they are not able to get the raw data.
HARRIS: Academics have often accepted that. For one thing many lab scientists don't have the expertise to do their own statistical analysis and drug companies do have scientists with those talents on staff. But as flaps like this one show, that division of labor can be quite problematic. Just how data are analyzed can affect a study's conclusion and the company's profit. And the choice of statistical method itself may determine whether a drug comes out looking safe or unsafe. As a result of this growing concern, the Association of American Medical Colleges has recently issued guidelines for conducting drug studies. The organization recommends that for major studies at least the academics not the company should conduct the data analysis. Richard Harris, NPR News, Washington.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.