Here are the outlines of the issue, as I understand it:
- In 2004, Dr. Douglas Kiel, a researcher from Hebrew Senior Life - a terrific nursing home and research institution affiliated with Harvard (also not so shabby a place) - were conducting a study of hip protectors. At the time, evidence that hip-protectors prevented hip fractures was unclear. Previous studies were vulnerable to confounding - those in the group who wore hip protectors may have been different in important ways from those who did not wear hip protectors. Those differences, rather than the hip protectors, may have led to findings that hip proctors protect against fracture. Dr. Kiel and colleagues came up with the novel idea of allowing patients (nursing home residents) to serve as their own controls: they would put a hip protector on only one hip, and the other hip would serve as the "control." There would be no differences between study subjects because each subject experienced both the intervention and control at the same time! The NIH funded the study at a cost of 8.5 million dollars.
- As the study progressed, internal documents (emails) show that researchers began to worry that something unexpected was happening. More study subjects were falling toward the side of the hip protector than the other side. They were more likely to experience fractures on the side with the hip protector. These findings were not disclosed to study subjects.
- Fortunately, because this was an intervention trial that posed potential serious risks to human subjects (hip fracture), a Data Safety Monitoring Board was established to review the study data and protect the health of study subjects.
- Unfortunately, researchers decided not to disclose the extent of the problem to the Data Safety Monitoring Board. They made a conscious effort not to draw attention to it, because they feared the study would be shut down.
- The results of the trail were published in JAMA in 2007. The study demonstrated that hip protectors did not protect against hip fractures. The accompanying editorial picked up on the finding that more patients experienced hip fracture on the side with the protector (13 fractures) than the side without (7 fractures), and concluded that either hip protectors are ineffective and cause fractures, or, "the use of a 1-sided hip protector modified the individual’s standing and walking, such that it increased the risk of hazardous fall onto the protected hip."
- Post-hoc review of pilot data used to establish the feasibility of this large trial subsequently revealed that there was clear evidence that hip protectors worn on one side of the body led to falls toward that side. Here is the data:
Fell to left
Fell to right
Pad on left
Pad on right
- In 2008, HipSaver's Inc sued Dr. Kiel for unfairly disparaging hip protectors. The law suit led to uncovering the internal documents described above, and subsequent investigations by the Federal Government (HHS), JAMA, and Hebrew Senior Life.
- HHS asked the institutions involved in the study to contact study participants and disclose the risks to which they had been exposed.
Investigations are ongoing, so it may be that information comes to light that is less damning to the study investigators. But from what I've read, it seems they failed to provide information about known risks of their intervention to participants, and hid this information from the Data Safety Monitoring Board and JAMA.
What's fascinating to me about this study is the disregard for the welfare of human subjects by the investigators involved. In previous posts I've spoken out against the dangers of industry sponsorship of trails. This fiasco demonstrates the dangers of being wedded to science, without sufficient concern for the dangers posed to study participants - people.
This fiasco will undoubtedly have fall out for other researchers, in terms of increased scrutiny of research protocols, red tape, IRB oversight, etc. As it should. Ethical violations of this sort are a serious matter. It's a shame, though, because in the end all those extra hoops to jump through will protect some patients, but also create barriers that may prevent other great studies from being conducted. And ultimately that's not in everyone's best interest either.
by: Alex Smith