The Seattle Times recently ran an article with an interesting contrast to Facebook’s no-information, no-consent experiment on news feeds. The writer discussed a Seattle hospital’s plans to conduct an experiment on ER patients without necessarily telling them. The experiment will test the effectiveness of a drug solution on bleeding around the brain. Once the study starts (sometime in the next few months), anyone arriving at Harborview Medical Center’s ER with a head injury could be a part of it. The patients may not be conscious, and the doctors may not have time to conduct the lengthy informed consent process that’s normally required for medical experiments. To learn whether the treatment works means skipping that core tenet of research on human subjects.
Where is the tie-in to the Facebook study? At first glance, Harborview’s plan resembles the much-publicized Facebook move, but with more serious consequences. Facebook didn’t tell any users that it was manipulating news feeds to see how we would react, and at Harborview the ER staff could administer a drug with an experimental use without notifying us while we’re possibly helpless. What is the deletion of happy or unhappy language compared to that?
One difference in Harborview’s case is that somebody’s watching. A whole host of somebodies, actually. Some of this study’s funding comes from federal agencies, which means it must meet their ethical guidelines. And the research is happening at the hospital of a public university, so state and school jurisdictions are additional layers of potential oversight. Each level has requirements for ensuring any research is ethical – coming down to the researchers and the university’s ethics review board. To conduct research in the emergency room, the ethical path is narrow and restrictive. In comparison, Facebook only had an internal vetting process.
The prospect of experimenting on someone without explicit permission sets off all kinds of alarms, but in the case of the emergency room, it also could be the only way to find out what works. This experiment is testing the effectiveness of Tranxemia, a solution that is FDA-approved for other situations. Tranxemia has a solid history of doing better than other treatments in reducing blood loss during heart surgery, dental surgery, and difficult deliveries. Harborview (and others) want to see whether Tranxemia can do the same for bleeding in the brain.
How do we learn the best treatments without comparisons? Can we risk someone’s life to find out? What if that risk could save future lives, or even the life of the person involved? These questions are hard, the risks high, but the potential benefits are life-saving. A few years ago, another ER study tested several methods of CPR and illustrated which ones helped the most.
To allow this kind of study, regulators and ethicists face extensive requirements. In summary:
- They need to be certain the proposed method is the only way to test the question
- They need to be certain the study sponsor will watch the results closely
- In lieu of seeking consent before treatment, they require an outreach program to the community
Harborview has allocated $34,000 to conduct a survey, make announcements, issue flyers, and hold meetings. Since bicyclists are some of the most likely to visit the ER for head injuries, the hospital is teaming up with a local bike club to publicize the experiment. If someone be a part of the study, the hospital will provide a bracelet. One could argue whether or not these measures are adequate notification to the community, and that argument is the ethics committee’s job.
For me, the fundamental difference between Harborview’s approach and Facebook’s is respect. “Respect for persons” has been a fundamental element of human research ethics for over 35 years, and the regulations codify ways researchers can honor it. The guidance for emergency room research is strict because not obtaining prior consent suggests a lack of respect. To balance that out, the hospital must follow all of the necessary steps:
- Prove there’s a lack of alternatives
- Demonstrate a clear possibility of benefit
- Let the community know
- And an independent group—one with members who are not beholden to the hospital—has to agree with the plan
And what about Facebook? The social media provider tweaked our news feeds without telling us. When compared with treating head injuries, the stakes with Facebook are nowhere near as high. But Facebook deliberately manipulated information to see how people would react, and did nothing to alert people that they were being manipulated, let alone asking their permission. To me, nothing about that demonstrates respect.