Verified by 
In a recent essay I asked, "Should Animals Be Used To Test Party Pills?" I concluded that the answer is "No" and noted there are many different views on whether or not nonhuman animals (animals) should be used to test drugs that are meant to be used solely or primarily by human animals. And, there's a good reason for debate and skepticism.
It's well known that the vast majority of drugs that pass tests on animals do not work on humans. To quote from this report, "The FDA reports that 92 percent of drugs approved for testing in humans fail to receive approval for human use. This failure rate has increased from 86 percent in 1985, in spite of all the 'advances and refinements' intended to make animal tests more accurate." If we look at these data in another way, only about 8% of drugs that pass tests on nonhuman animals also supposedly work on humans. I know that if someone told me I only had an 8% chance of leaving my home and getting to town I'd decide not to venture out.
Hundreds of thousands of people die annually from drugs that work on animals
Now, a recent essay in New Scientist magazine by research scientists Kathy Archibald and Robert Coleman called "How human biology can prevent drug deaths" presents some updated data that lead the authors also to conclude that we can do significantly better to protect humans by not using animals to test drugs.
This essay begins, "ADVERSE drug reactions are a major cause of death, killing 197,000 people annually in the European Union and upwards of 100,000 in the US. Little coverage is given to such grim statistics by governments or pharmaceutical companies, so patients and their doctors are not primed to be as vigilant as they should be, and adverse drug reactions (ADRs) remain seriously under-recognised and under-reported." (my emphasis)
The authors reiterate the 92% failure rate because of drug toxicity, and also note, "in 2006, six people enrolled in a UK trial of the drug TGN1412 were hospitalised after developing multiple organ failure. Many clinical trials are now conducted in India, where, according to India's Tribune newspaper, at least 1725 people died in drug trials between 2007 and 2011. Clearly, there is an urgent need for better methods to predict the safety of medicines for patients as well as volunteers in clinical trials." Furthermore, a recent study showed that "animal tests missed 81 per cent of the serious side effects of 43 drugs that went on to harm patients."
Researchers Archibald and Coleman wonder why governments and drug companies continue to use and to defend their incredibly poor records using animal tests while there are new technologies readily available that will result in increased patient safety and decreased use of animals. Indeed, the organization Safer Medicines of which Kathy Archibald is the director has been told by the UK Department of Health and the prime minister that "human biology-based tests are not better able to predict adverse drug reactions in humans than animal tests".
While the United States National Research Council has recognized the need to replace animal tests with "more efficient in vitro tests and computational techniques" there's clearly a major problem in the UK where governments and drug companies deny there's a significant problem with animal testing.
If governments and drug companies paid attention to the data it would be a win-win situation for all of the animals, human and nonhuman
The New Scientist essay to which I'm referring is a very important one because the conclusion that animal tests do not work is not merely being put forth by people advocating for animal protection, but also by respected scientists who want to see drugs do what they're purported to do for humans who need them and who also care about the animals who are unnecessarily killed because of the stubborn denial of the inadequacy of animal tests by those who continue to advocate for them in light of strongly contradictory data. If governments and drug companies paid attention to the data it would be a win-win situation for all the animals, human and nonhuman.
