A risk-benefit matrix should ask the right questions

Here is how to develop an effective tool

Simply asking investigators to provide a thorough discussion of a proposed study’s probability of risk and potential benefits may not always result in an effective analysis that can be used by IRBs to assess a protocol. This is why it’s a good strategy to ask investigators to use a risk-benefit tool, such as a matrix, in analyzing their protocols.

"The burden is on investigators to demonstrate what the risks of harm are," says Paul B. Gold, PhD, an assistant professor of psychiatry at the Medical University of South Carolina in Charleston. He also is an associate vice chair for one of the university’s three IRBs, and has been a principal investigator for studies in which he has submitted to an IRB his own risk-benefit matrix as part of the protocols.

Research problems have been highlighted in the popular media so often in recent years that it’s incumbent upon investigators and IRBs to help reverse the trend of low respect for medical research, he says. "We have a task in front of us, and we do have to gain credibility," Gold says. "Through being transparent and showing what we do and being honest with what we know and what we don’t know, we’ll get the credibility back."

With this objective in mind, Gold has developed a risk-benefit matrix for harm assessment. This matrix forces an investigator to take a clear and honest look at the proposed study’s probability and magnitude of harm, and it provides a blueprint for information that will be disclosed to potential subjects in the informed consent discussion and document. It also makes it very easy for an IRB to assess a protocol in terms of the balance between risks and benefits.

"When I set up a matrix for harm assessment, I’m also thinking statistically," Gold explains. "So I’m looking at things from four perspectives simultaneously: One is the ethical approach, another is statistical, third is from the perspective of the provider of care, and the fourth is as a patient." Gold then asks himself what a federal scientific reviewer would think about this protocol and whether an IRB would be comfortable approving it.

Here’s how Gold developed the matrix:

• Keep in mind the three major ethical principles: "In my matrix for beneficence, I have an assessment of risks and benefits," Gold explains. The three major principles are:

  1. Respect for people;
  2. Beneficence;
  3. Social justice.

• Understand the various types of benefits: In assessing a protocol in terms of the ethical principles, it’s important to consider that there are three levels of benefits, including direct benefit, collateral benefit, and societal benefit, views taken by the National Bioethics Advisory Commission and Nancy M.P. King, JD, professor of social medicine at the University of North Carolina-Chapel Hill.

A particular study may have all three benefits, or perhaps only one of the three. But if the investigator hasn’t given this area some thought, it’s possible that both the investigator and subjects will misunderstand the potential benefits. "Direct benefit means the individual participant will gain personally from participating in the study," Gold says. "That’s a rigorous definition and extremely difficult to meet and hard to estimate."

A collateral or indirect benefit is a benefit that occurs as a result of the person’s participation in a study, but it was not the goal or intention of the study to produce this benefit. The societal benefit may include solving a problem therapeutically or increasing the scientific community’s knowledge of a process in hopes that this information will one day lead to developing a new therapy.

• Estimate probability of direct benefit: Before researchers can say that a particular study may provide a direct benefit to participants, they should analyze the probability and magnitude of such an outcome, he says. "The direct benefit generally is going to be very low for the individual, but that doesn’t nullify the value of a study," Gold says. "That just means participants should be aware of that fact, so they don’t fall victim to the therapeutic misconception."

It’s also important for investigators to be aware of their own therapeutic misconceptions, especially if they recruit volunteers from their clinical case loads and take on the dual role of provider and investigator, he says. As an example, Gold has calculated that a particular study he reviewed for his IRB would have a one-in-eight likelihood of direct benefit, based on a randomization scheme, the study’s anticipated results, and other information included in the protocol. "I was making a guess, and we do that all the time anyway," he says. "Just like with harms, the probability and magnitude of benefits should be determined, even though it is tougher with a benefit."

• Provide practical applications analysis: Gold’s matrix includes three practical applications, which are:

  1. Informed consent;
  2. Assessment of risks and benefits;
  3. Selection of subjects.

The informed consent discussion, for example, would include information about the study, an attempt to make certain the information easily can be understood, and an effort to show how subjects volunteered without any coercion.

• Look at risks of all facets of harm: Gold typically will begin to assess a study with regard to all of the possible harms, including psychological, social, economic, and physiological. "Physiological gets most of the attention because those harms often carry a probability for severe injury, possibly even death," he says, "whereas the others may carry humiliation, loss of job, or loss of insurance."

For instance, suppose an insurance company learns from a research study that a particular client suffers from clinical depression and has a problem with alcohol abuse. "The insurance company may say that the person drinks alcohol, which is a voluntary behavior, and therefore the company won’t insure the person for major medical problems," Gold says. "Researchers take care to preserve confidentiality, but there still are potential harms, and I lay these out along with specific harms, such as psychological discomfort in an interview." Economic harms might include job loss and other things that occur in people’s daily lives, but are accentuated because of a study.

• Predict harm probability and magnitude: With each potential harm identified for a particular study, Gold will look try to determine the probability or likelihood of this harm occurring and the magnitude or how damaging it would be if it did occur. "You may have the probability of one in 10,000 for a harmful reaction to a medication, and that may not require as much attention to protect research volunteers as the probability of one in 10 that someone’s HIV status gets communicated to the wrong people," Gold explains. "So that [latter] harm needs more attention so you can protect subjects against the breach of confidentiality."

"In my matrix of harm, probability, and magnitude, I also have a column of what we’re going to do to minimize a harm," Gold adds. As part of the process of completing a harm-benefit matrix, investigators should list all of the measures they will take to minimize the potential for harm. These may include the following:

  • How information will be collected, stored, and protected.
  • How closely adverse events will be monitored and the likelihood of these occurring.
  • How frequently subjects will be contacted.
  • Whether participants have mechanisms for contacting investigators.
  • Whether participants will be given written explanations of the potential adverse events, and whether they can remember what these are.
  • Who is going to be in charge of coordinating this investigator-subject communication.

"I find it hard to get that level of specificity from investigators submitting protocols to the IRB, but it’s necessary," Gold says. "Having someone volunteer in your study is a privilege and not somebody’s right." Volunteers are heroes, Gold adds. "If they were perceived as heroes for going into research, we’d all be very careful."