In lieu of an abstract, here is a brief excerpt of the content:

  • Learning from Clinical Experience
  • Mary Ann Baily (bio)

In this issue of the Report, James Sabin and colleagues consider the ethics of conducting randomized trials in health plans to compare the effectiveness of drugs approved by the Food and Drug Administration that are considered to be in clinical equipoise—there is no consensus among experts that one drug is superior to the others. The authors argue that health plans should publicize the fact that they allow comparative effectiveness trials to be run, but the plans do not have to obtain the individual informed consent of patients. The article's most important message is that we should be using everyday medical practice as a source of systematic information on treatment effectiveness.

Some people who are deeply committed to the protection of the rights of human research subjects find just the thought of conducting a randomized clinical trial without consent profoundly shocking. They should read the article before dismissing cluster randomized trials (CRTs) outright. I admit that I came to the article already believing that we have to make it easier to develop information on effectiveness. Nevertheless, I think readers will find that the authors make a carefully reasoned and ultimately convincing case for the ethical acceptability of their approach. They also report interesting information about attitudes toward CRTs, gleaned from a two-hour discussion of the topic by the ethics advisory group of Harvard Pilgrim Health Care, and over one hundred interviews with patients, clinicians, and health plan leaders.

Their approach won't be easy to implement, however. The health plans have to find drugs that are in equipoise, and the results will be better if it is the right kind of equipoise—if the lack of consensus among experts means that most physicians are undecided about which drug is superior (in "internal equipoise") rather than committed to the belief that one drug is much better than the others. The trial must be carefully structured to meet statistical requirements. And finally, the health plans must deal with the attitudes revealed in the empirical work. Most of the patients interviewed had great difficulty with the concept of equipoise. They were unaware of the fundamental uncertainty that pervades medical care and didn't understand how physicians could not know which drug was best. They were also very wary of being included in research. The health plan leaders interviewed worried that patient concerns about being research subjects would affect plan enrollment. They emphasized the novelty of CRT and the need to educate patients and other stakeholders about it.

The authors do a good job of explaining why we can't rely completely on classic randomized controlled trials for information on effectiveness. They describe the variation in everyday practice and explain how it helps to justify the CRT approach. Since randomization makes CRTs controversial, it's worth noting that there are other ways to learn from clinical experience that do not involve randomization. The randomized trial is not the most common approach used in social science research. Instead, social scientists use a wide variety of statistical methods to learn from data collected on what happens in everyday life. Also, business managers routinely use data on their own operations to optimize organizational performance and cost. The same reasoning used to argue for the ethical acceptability of CRT can be used to justify the use of these methods in health care.

The health care system produces large amounts of data on treatments and outcomes, yet we have failed to take maximum advantage of it to make care more effective and efficient. Many factors contribute to this, but one is definitely patient concern about being used as research subjects. Patients fear that their care will be compromised and that they will be put at significant risk in pursuit of some vague future social benefit. Meanwhile, they don't realize how much uncertainty underlies the practice of medicine and how often clinical decisions have little basis in evidence. As a result, they don't appreciate how much their own care could be improved through systematic study of the natural experiments reflected in current treatment patterns. The only risk to patients would be from the data use, and this could...

pdf

Share