Matthew Herper’s article in Forbes, “Ebola, Ethics, And The New Normal Of Scary Germs” is good journalism. It’s centered around an article by Borio and Cox, in the New England Journal of Medicine, “…arguing that despite the hopes of some public health experts, both vaccines and drugs will need to be tested against control groups that include a placebo.”
This makes me mad, which is why the title of this post includes the word “manslaughter.” The purpose here is to sharpen up the logic surrounding the assessment of Borio/Cox with my own hostile appraisal. In the post “Ebola, Public Health, and Sloppy Thinking Part 1”, I wrote, “The bureaucracy of health and medicine, which, while providing the individual with paternal protection from naivete and quackery, has a knowledge base that, Dr. Ioannidis has shown, has severe methodological flaws.”
and, “The decision processes of the CDC and NIH are reliant on the same decision processes that create the studies torn apart by Ioannidis. It’s the same culture. The brightest minds in CDC and NIH must know this in an academic way. But if all the real estate, stretching to the horizon, is quicksand, what does a builder do? You build anyway, institutionalizing defective thinking.”
Medical ethics is joined at the hip to this culture. In the past, the requirements of medical ethics, as defined by individuals in the class of Borio/Cox, have been responsible for protocols that, rigidly adhered to, resulted in the effective denial of drugs to cancer patients. One of the gyrations of the medical establishment has been alternations in policies regarding risky or questionable drugs. The point to be extracted is not whether a particular drug should be offered or denied. It is that the reversals show that being a specialist in some branch of “ethics” isn’t all it’s cracked up to be.
Medical ethics exists as a specialty not because it answers ethical questions, but because it appears to do so. We might as well do the flip-flops ourselves. In this case, Borio/Cox assert that blind trials with control placebo groups are required to validate Ebola vaccines before they are deployed.
This they require, while in some countries, Ebola “hospitals” were recently equivalent to death camps. But Borio/Cox are correct about supportive care. The African overall mortality rate appears to be dropping; in October, it was about 2/3 overall, dropping in late November to about a third. It is likely that this is due to better supportive care. So why don’t Borio/Cox win the argument?
There is no objective solution to any question of medical ethics. It is inherently a subjective question. There are those of us who feel that “ethics and religion are opinions that you have.” I regret if any sensibilities are hurt, but our legal system tacitly recognizes this. Neither of these statements is acceptable in a court of law:
- “God made me do it.”
- “My ethics required that I terminate the victim’s life.”
In a statistical fashion, the members of the test group, including those receiving the placebo, have one of these fates:
- Death as a result of the vaccine.
- Death as a result of not receiving the vaccine.
- Survival as a result of the vaccine.
- Survival as a result of not receiving the vaccine.
- No effect on survival.
If the statistical element were absent, so that the will of the ethical experimenter directly determined the fate of individuals, it would be clear that the ethicist was, to the individual subject, in every way equivalent to the Nazi Death Camp guard selecting the unfortunates for the gas chambers.
Does statistics make it alright? I don’t think so. But you argue, we really don’t know whether the vaccine works at all, or whether it could harm. There are more than a few problematic pathogens. Dengue is an example. There are four strains of dengue. While an infection with one strain provides immunity against that strain, it makes subsequent infections worse. By the time you reach the fourth strain, you might be dead. AIDS is another. Other errors are choosing the wrong antigen as as target, or a negative effect from immunization against the wrong viral strain. With bacterial pathogens, there can be cross-reactivity with human tissues, resulting in an autoimmune disorder.
That’s what separates the experiment from murder, and makes it a laudable activity. That could be true, if human experiment were the only tool available. Fortunately, there are other ways of knowing. A common element of vaccine development is the animal model, a species with response to infection similar to the human. The symptomology and immune response of the model animal is extensively characterized, so that it can serve as a proxy for the human in preclinical studies. For influenza, the animal of choice is the ferret. It’s small, cheap, and it catches the flu very easily.
The FDA deals with the current Ebola scenario with the “animal rule” adopted in 2002. Quoting from the The NIH article, “Current Ebola vaccines”,
“In these cases the ”animal rule” permits efficacy data from studies using animal models that accurately recapitulate human disease…However, for this purpose the correlate(s) of protection predictive of survival in the relevant animal model have to be defined, so that these data can be used to predict human efficacy.”
The animal in question is any nonhuman primate, although the Great Apes, which in this case practically means chimpanzee, confer some advantage. In these animals, Ebola is typically 100% fatal, compared to the human mortality, which maxes out at 90. The caveat comes in the next paragraph:
“Unfortunately, until now there has been only limited success in defining these correlates of immunity, although it could be shown that there is some correlation between antibody titres (but not necessarily neutralizing antibody titres) and survival…”
So there is a rule, and there is math, and it is likely that the conclusion of Borio/Cox is due to the failure of the animal model to satisfy the math of the rule. By how much, I wonder? Sadly, I do not have access to the New England Journal of Medicine. This excruciatingly vital part of the debate is hidden behind a paywall, an archaic aspect of scientific publishing that some of us think goes against academic freedom and the right of the public to know.
The referral to “correlates of immunity” suggests that the rule may require more than survival rates to validate the vaccine. Correlates are measurements of antibody concentrations and immune system activation. If this is the case, if the rule requires more than survival rates of the model animal, it expresses a need to quantize the “why” of survival. Quantization is not the same as understanding, but this is a bureaucracy we have to feed. They want to know “why it works”, because they want to assume that the numbers will be transferable to human responses.
But do we always need to know why? Suppose you’re trapped in a submarine and water is streaming in. Someone taps out a message in Morse Code on the hull: “PULL LEVER B TO SURFACE.” Do you need to know why? Just pull the lever.
Borio and Cox suggest that the ethically correct course is to build lots of intensive care wards. But it’s not simply a case of hooking up the right I/V lines to rehydrate. In the industrialized world, intensive care units have climate control. In Africa, isolation wards are tin-roofed structures or tents, with ventilation restricted by the need for isolation. Building modern facilities with all the features favorable to patient survival would be the Manhattan Project of Africa. And the supporting infrastructure doesn’t exist. How many people would die in the interim?
Given that Borio/Cox are part and parcel of the system of thought that has so far bungled handling of the Ebola epidemic, a very high standard of thought, exhibiting truly independent reasoning, is required. My suspicion is that their points of view suffer institutional contamination. Another way of putting it would be that they are defending their intellectual turf. Perhaps they would refer to the Hippocratic Oath: Do no harm. But Hippocrates saw one patient at a time. Over what time period? Over what sample? What makes their requirement different from manslaughter? I want it technical.
A reductio ad absurdum drives it home. Suppose a vaccine is developed against malaria, but it has a mortality somewhat higher than the yellow fever vaccine, which, containing a live, attenuated virus, actually kills about 1 in 50,000 recipients. Borio/Cox could say (putting ridiculous words in their mouths), “You really ought to get rid of the mosquito.”
Yes, it is ridiculous. Yet, having come full circle, we see that murder and sainthood are neighbors. They come out of their abodes, mosey over to the fence, slap high fives, shoot the bull, and get along just fine.