87% of the studies failed to randomize the selection of animals (a technique used to reduce "selection bias") and 86% did not use "blinding" (another technique to reduce researcher bias).
Also, "only 59% of the studies stated the hypothesis or objective of the study and the number and characteristics of the animals used." A 2017 study found further flaws in animal studies including "incorrect data interpretation, unforeseen technical issues, incorrectly constituted (or absent) control groups, selective data reporting, inadequate or varying software systems, and blatant fraud." Since the majority of animals used in biomedical research are killed during or after the experiments, and since many suffer during the studies, the lives and wellbeing of animals are routinely sacrificed for poor research.
Microfluidic chips ("organs on a chip"), which are lined with human cells and recreate the functions of human organs, are in advanced stages of development.
Computer models, such as virtual reconstructions of human molecular structures, can predict the toxicity of substances without invasive experiments on animals.
According to Humane Society International, animals used in experiments are commonly subjected to force feeding, forced inhalation, food and water deprivation, prolonged periods of physical restraint, the infliction of burns and other wounds to study the healing process, the infliction of pain to study its effects and remedies, and "killing by carbon dioxide asphyxiation, neck-breaking, decapitation, or other means." The Draize eye test, used by cosmetics companies to evaluate irritation caused by shampoos and other products, involves rabbits being incapacitated in stocks with their eyelids held open by clips, sometimes for multiple days, so they cannot blink away the products being tested.
The US Department of Agriculture (USDA) reported in 2016 that 71,370 animals suffered pain during experiments while being given no anesthesia for relief, including 1,272 nonhuman primates, 5,771 rabbits, 24,566 guinea pigs, and 33,280 hamsters.
Humane treatment is enforced by each facility's IACUC, and most major research institutions' programs are voluntarily reviewed for humane practices by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC).
Laboratory mice, for example, live for only two to three years, so researchers can study the effects of treatments or genetic manipulation over a whole lifespan, or across several generations, which would be infeasible using human subjects.
stated that "The low predictivity of animal experiments in research areas allowing direct comparisons of mouse versus human data puts strong doubt on the usefulness of animal data as key technology to predict human safety." Biotechnology company Empiriko invented synthetic livers which can predict the liver's metabolic reactions to drugs in a process that is quicker, cheaper, and more accurate than animal testing; in one trial it provided a level of specificity which previously would have required testing on 1,000 rats and 100 dogs.
A peer-reviewed study found serious flaws in the majority of publicly funded US and UK animal studies using rodents and primates.