Determining risk-benefit of clinical trial requires careful, statistical thinking
Formula shows true, not imagined risk
Investigators weighing the risks and benefits of a randomized clinical trial might try using a new statistical concept that offers consistency and clarity to their decision-making process.
Statistical thinking can provide a solid and more objective basis for describing risks when a principal investigator submits a protocol to an IRB.
"The risk has implications for informed consent and how this is viewed by the IRB," says Martin L. Lesser, PhD, director, investigator, and clinical professor at the Feinstein Institute for Medical Research in Manhasset, NY. Lesser also is chairman of the institute's IRB.
"I wanted to compare the risk-benefit ratio for participating in the study with the risk-benefit ratio if you used the same subjects, but not in the study," Lesser explains. "If the risk is the same then the risk is not any greater than those not in the study, and that means it's not greater than minimal risk."
The model only works in the case where the two or three treatments in the clinical trial are accepted as standard therapy, he notes.
"It only applies to non-investigational therapy," he says.
From a statistical point of view, people use some kind of risk probability model whenever they make decisions. The risk might be the loss of a reward, for example, Lesser says.
"We make decisions about this every day," he says. "So we weigh the potential risks and side effects against potential benefits and quality of life."
If the risk-benefit ratio is low enough then an IRB might determine the protocol is acceptable and approve it. Then participants, also weighing risks-benefits, will decide to participate in the study.
"We want the risk to be acceptable, and we each have an acceptability limit," Lesser says.
Statisticians involved in this process look at the total risk, weighing the risk of taking one action versus the probability of risk if another action was taken, he says.
A statistician wants to know whether the risk-benefit ratio is similar to, equal to, or different from what it would be if the potential clinical trial participant was a regular old patient who is not involved in the trial, Lesser says.
Lesser and co-investigators came up with this statistical concept during the study of risk in a randomized clinical trial about glucose control in the intensive care unit (ICU).1
"There were two things researchers wanted to compare," Lesser explains. "One was the sliding scale insulin where you check somebody's glucose every four hours, and based on the results you adjust the level."
The second method was to use insulin glargine, a well-known type of long-lasting insulin that diabetics use over time, he says.
Neither treatment is experimental, and diabetic patients frequently use both of these methods.
"Investigators wanted patients to receive one or both of these to see if using glargine was beneficial," Lesser says. "The outcome would be some measure of glucose control."
Lesser examined the question of how to decide whether this study would constitute a minimum risk clinical trial.
"I thought the definition of minimal risk was risk encountered in a study that was no greater than that encountered in daily life," Lesser says. "I broaden that concept in my mind so that it's not just the risk encountered in daily life, but risk encountered in daily life with your disease."
So if a patient with cancer goes in for chemotherapy, then this is standard care, he adds.
For example, in the clinical treatment of diabetes in the ICU, whether treatment involves using sliding scale insulin with or without glargine is up to the physician's discretion.
"So if I am a patient going into the ICU, my treatment depends on whether my doctor believes in using sliding scale insulin only or sliding scale with glargine," Lesser says.
Lesser and co-investigators assigned a formula depicting this risk and then compared it to the risk of being randomized into one arm or the other in a clinical trial setting.
"So now if I want to prove that the risk of participating in the study is no greater or not substantially greater than being treated in the community outside of the study, then I have to compare the risk-benefit ratio of the study to the risk-benefit ratio of the community treatment," Lesser explains. "If RBR is less than RBRc, then the risk of participating in the randomized trial is no greater than receiving randomized treatment and it's minimal risk."
The point of using a statistical concept to evaluate risk is to provide an alternative to pure IRB member subjectivity on the risk determination.
"My argument is the study might be considered minimal risk if you can show that the risk of participating in the randomized trial is not much different than the risk of being treated in the general community if you had this disease," Lesser says.
Reference:
- Lesser ML, Kohn N, Neal EV. Using statistical concepts to determine risk level of randomized clinical trials that compare two non-investigational therapies. Poster submitted to the 2008 PRIM&R Advancing Ethical Research Conference, held Nov. 17-19, 2008, in Orlando, FL.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content