The recently approved experiments at the University of Wisconsin-Madison involving rhesus monkey “peer-rearing”, a form of early life stress, have raised concerns among some academic ethicists. In particular, Lori Gruen, a Professor of Philosophy at Wesleyan University, has posted two blog entries about the study arguing that the study “is not ethically defensible” on the grounds it violates the spirit of the Animal Welfare Act, the model is not appropriate, and there are better ways to achieve the same benefit. She also criticizes the approval process for the protocol.
I want to thank Professor Gruen for her attention to this issue. Regardless of whether one supports or opposes the research, we can agree that it raises important ethical issues that deserve an open and informed discussion. Here, I wish to further that discussion by responding to Gruen’s concerns.
But first, I need to point out that Gruen’s writing implies that the effect of peer-rearing on the well-being of these monkeys is more severe than it really is. In her second post, Gruen states:
“There was no oversight system in place back in the days when Harry Harlow’s experiments psychologically tormenting baby monkeys were making news. Surely that sort of horrible work in which infant primates are taken from their mothers to make them crazy wouldn’t be approved of today. On my recent visit to the University of Wisconsin I was shocked to learn otherwise.”
In fact, the proposed studies produce at most a moderate early life stress. I have described the peer-rearing experiment design elsewhere, but will restate two pertinent observations. First, peer-rearing actually is used to save a monkey’s life when a mother rhesus rejects its infant and a foster mother cannot be found. Second, clinical records show that, in the UW-Madison colony, peer-rearing typically only is associated with thumb sucking behavior (just one young monkey in 10 years injured herself, and she recovered). Peer-rearing will measurably increase anxiety, as required by the experimental design, but does not make a monkey “crazy”.
Gruen suggests this study “violates the spirit, if not the promulgated regulations, of the Animal welfare Act (AWA) which specifically requires that the psychological well-being of primates be protected (not intentionally destroyed).” Later, however, Gruen concedes that “the AWA allows exceptions for scientific reasons if they are approved by an Institutional Animal Care and Use Committee (IACUC).” The purpose of peer-rearing in this study is to create monkeys that are environmentally predisposed to anxiety, so that we can compare their brain chemistry and pattern of gene expression to monkeys raised with their mothers. Thus, peer-rearing in this study both has a clear scientific rationale, and it was IACUC approved. Its use therefore violates neither spirit nor letter of the law.
Next, Gruen states, “it isn’t clear that the monkey model is adequate.” In support of this claim, she suggests, with unclear relevance, that differences in the developmental onset of nerve myelination between humans and chimpanzees makes the rhesus monkey a bad model for humans. To the contrary, at the level of behavior and brain neural pathways, monkeys and humans manifest anxiety very similarly. Monkeys are considered to be a scientifically appropriate species for these studies.
To further support her contention, Gruen then quotes a pediatric psychologist who says that she
“sees many patients in our offices who have grown up under ‘adverse’ conditions, and they are hardly a uniform set. There are a wide variety of stressors that traumatized human children have to deal with, including various types of abuse and neglect, and these can never be replicated with any kind of accuracy using animal models.”
I accept this statement, but rather than indicating that the monkey is a bad model, it points out why humans would be a terrible model even if they could be ethically used. There is extreme variability in both the causes of anxiety in humans and the quality of the person’s subsequent environment. Peer-rearing, in contrast, produces a moderate adversity that reliably and more uniformly produces brain changes associated with anxiety, and thus permits comparison between groups of individuals who differ principally in this one variable of early experience—peer-rearing versus maternal rearing. This ability to carefully match control and experimental animals is critical to this experimental design, and makes it likely that any differences found between groups will be highly relevant to the root causes of anxiety in monkeys. The similarities of anxiety between monkeys and humans suggest this study also will be highly relevant to the causes of anxiety in humans. In short, the monkey IS a good model.
So far, I’ve argued that the peer-rearing study is legal and likely to provide valuable information. Now let’s look at whether there are better ways to accomplish the same objectives. The goal of the current study is to identify changes in brain pathway chemistry and gene expression in anxious monkeys. What does Gruen suggest as an alternative?
She proposes that the research funds be redirected to social programs that “can prevent the psychological harms that arise from childhood trauma.” But when dealing with complex problems, multiple approaches are almost always more effective than putting all of our eggs in one basket. By all means, let’s spend money supporting parenting skills classes, affordable childcare, and so forth. But that approach will never eliminate the problem or reach everyone in need. Given the cost of social programs, redirecting all the research funding from these studies would barely make a dent.
A more significant difficulty with Gruen’s argument is the fact that the purpose of these experiments is not to eliminate anxiety, but rather to identify molecular changes in the brain. This more modest and specific goal can lead to targets for medical treatments. And it can’t be achieved with social programs. Remove the funding, and it would have to be dropped.
So to achieve the benefit of reducing anxiety disorders, let’s tackle the problem from several directions. Fund social programs. But when that approach fails to prevent disease, let’s supplement it with effective treatments that only the research can give us.
I have explained why I believe the research is legal, uses an appropriate model, and proposes methods that are necessary to achieving its stated aim. Equally important is the question of whether the IACUC followed required procedures to reach a similar conclusion.
IACUCs are mandated by law, and, among other responsibilities, must review any proposed use of animals in research, teaching, or outreach before it can begin. Each IACUC must have at least a veterinarian with expertise in the species used, a scientist knowledgeable about experimental design, a non-scientist, and a community member not affiliated with the institution. A typical proposal (the “protocol”) undergoing full IACUC review will be assigned to a single committee, read beforehand, discussed, and voted up or down. I calculate that review of a relatively uncomplicated protocol not associated with significant animal pain or distress takes around 4 person-hours (number of reviewers multiplied by time of review).
What about the quality of IACUC review? Gruen gives two reasons for thinking it must have been poor. First, she states, “the oversight committee chairs [at UW-Madison] told me they have never rejected a proposal. Not one.” I expect she considers failure to reject a protocol to imply that the review process is a rubber stamp. Let’s dig deeper into the facts.
When our IAUCUC chairs say “failure to reject”, they mean that almost all protocols eventually reach a state that the IACUC can accept. There are two principle reasons for this. First, protocol preparation is an intensive process, involving careful description by investigators of what they want to do and how they want to do it. They typically receive extensive veterinary input. In other words, the protocols are good to begin with. Second, even after careful preparation, very few protocols are approved the first time around. Data from IACUC protocol review during January 2013 (a typical month) indicates that, of 51 new or renewal protocols submitted, only 4 (8 percent) were approved as written. The rest were rejected, deferred, or “required modifications to secure approval”. Perhaps a more accurate statement would be that our IACUCs reject 92 percent of the proposals they receive. I do not believe it is a criticism of our system that ICUCs are willing to work with investigators until a protocol finally receives approval.
Second, Gruen quotes from an IACUC meeting minutes, in which members seemed concerned about the research and confused about their roles. The meeting she cites was the first time this committee had seen the protocol, and the fact that they were willing to openly express their concerns about this unfamiliar and ethically charged proposal is the sign of a robust review process. The quote from a veterinarian (who, by the way, was not the IACUC chair) implying uncertainty about IACUC authority was taken out of context. His intent was to indicate that IACUCs should not reject NIH-approved studies only because they have the potential to cause animal distress. All IACUC members understand that the committee can vote down any proposal.
For this peer-rearing study, the protocol was assigned to two IACUCs, since it requested use of animals from two facilities overseen by different committees. It was reviewed at five individual IACUC meetings and 3 meetings involving both committees. A conservative estimate of the total person-hours of review: 170. Over the course of review, the investigators were required to address adequately a total of 187 IACUC questions before they finally secured protocol approval. I have not seen a more intensive protocol review in 18 years as an IACUC member. Even those who disagree with the IACUC’s decision should acknowledge that, in this case, the data demonstrate that approval was nothing like a rubber stamp.
Gruen concludes her second blog, which primarily criticizes IACUCs, with the following question: “What does this say about the ethicists who serve on these oversight committees?” I’ll answer that. A faculty ethicist chaired one of the two IACUCs evaluating this research, and he also chaired the three combined IACUC meetings. As chair, he ensured that all components of ethical protocol review—identifying costs and benefits, considering alternatives and consequences—were fully and explicitly addressed.
When the final vote was taken, the protocol was approved by a majority vote of each IACUC. He may or may not have voted to approve (the final ballot was confidential), but he made certain that the process honored the spirit and the letter of the law. This kind of direct experience with the subject of one’s ethical enquiries is less common than it should be. I commend him and all the members of our IACUCs for their lengthy and thoughtful participation in the review of this and other important research.