The "completely randomised" and the "randomised block" are the only experimental designs suitable for widespread use in pre-clinical research.

Too many pre-clinical experiments are giving results which cannot be reproduced. This may be because the experiments are incorrectly designed. In "Completely randomized" (CR) and "Randomised block" (RB) experimental designs, both the assignment of treatments to experimental subjects and the order in which the experiment is done, are randomly determined.

On determining sample size in experiments involving laboratory animals.

Scientists using laboratory animals are under increasing pressure to justify their sample sizes using a "power analysis". In this paper I review the three methods currently used to determine sample size: "tradition" or "common sense", the "resource equation" and the "power analysis". I explain how, using the "KISS" approach, scientists can make a provisional choice of sample size using any method, and then easily estimate the effect size likely to be detectable according to a power analysis.

Genetically Defined Strains in Drug Development and Toxicity Testing.

There is growing concern about the poor quality and lack of repeatability of many pre-clinical experiments involving laboratory animals. According to one estimate as much as $28 billion is wasted annually in the USA alone in such studies. A decade ago the FDA's "Critical path" white paper noted that "The traditional tools used to assess product safety-animal toxicology and outcomes from human studies-have changed little over many decades and have largely not benefited from recent gains in scientific knowledge.

Randomized block experimental designs can increase the power and reproducibility of laboratory animal experiments.

Randomized block experimental designs have been widely used in agricultural and industrial research for many decades. Usually they are more powerful, have higher external validity, are less subject to bias, and produce more reproducible results than the completely randomized designs typically used in research involving laboratory animals. Reproducibility can be further increased by using time as a blocking factor.

Evidence should trump intuition by preferring inbred strains to outbred stocks in preclinical research.

Inbred strains of mice such as C57BL and BALB/c are more widely used in published work than outbred stocks of mice such as ICR and CD-1. In contrast, outbred stocks of rats such as Wistar and Sprague-Dawley are more widely used than inbred strains such as F344 and LEW. The properties of inbred and outbred mice and rats are briefly reviewed, and it is concluded that, with some exceptions, there is a strong case for using inbred strains in most controlled experiments.

The design and statistical analysis of animal experiments: introduction to this issue.

Animal research has made major contributions to the health and welfare of humans and domestic animals. Immunization, first developed against rabies and anthrax by Pasteur using dogs, sheep, and rabbits, is now used to control many infectious diseases. The first drug, Salvarsan, was developed by Ehrlich using rabbits infected with the organism causing syphilis.

The extended statistical analysis of toxicity tests using standardised effect sizes (SESs): a comparison of nine published papers.

The safety of chemicals, drugs, novel foods and genetically modified crops is often tested using repeat-dose sub-acute toxicity tests in rats or mice. It is important to avoid misinterpretations of the results as these tests are used to help determine safe exposure levels in humans. Treated and control groups are compared for a range of haematological, biochemical and other biomarkers which may indicate tissue damage or other adverse effects.

Extending the statistical analysis and graphical presentation of toxicity test results using standardized effect sizes.

The results of repeat-dose toxicity tests are usually presented as tables of means and standard deviations (SDs), with an indication of statistical significance for each biomarker. Interpretation is based mainly on the pattern of statistical significance rather than the magnitude of any response. Multiple statistical testing of many biomarkers leads to false-positive results and, with the exception of growth data, few graphical methods for showing the results are available.

The effects of shipping on early pregnancy in laboratory rats.

Although rats in various stages of pregnancy are routinely shipped by vendors, the effects of shipping on pregnancy outcomes have not been reported. This study examined the effects of shipping rats 1 day after mating. Two outbred stocks, (Crl:CD(SD), Crl:WI(Han)) and one inbred strain (F344/Crl) of rats (n=300/strain) were mated in a vendor barrier room at 3-month intervals five times, and either shipped the next day (total time in transit ∼24 hr) or held in the room of origin until parturition.

Inbred strains should replace outbred stocks in toxicology, safety testing, and drug development.

Methods of toxicity testing, barely changed for several decades, need to be improved. One way forward would be to use a small battery of inbred strains instead of the single outbred stock currently used in toxicity screening. Inbred strains are more stable, more uniform, more repeatable, and better defined than outbred stocks.

Improving the design and analysis of animal experiments: a personal odyssey.

Everybody's career depends on many chance factors: the people one meets, the opportunities which are available, or the state of a scientific discipline. Mine is no exception. I started out in agriculture, obtained a PhD in quantitative genetics, and spent most of my career concerned with the use of animals in biomedical research.

Improving toxicity screening and drug development by using genetically defined strains.

According to the US Food and Drugs Administration (Food and Drug Administration (2004) Challenge and opportunity on the critical path to new medical products.) "The inability to better assess and predict product safety leads to failures during clinical development and, occasionally, after marketing". This increases the cost of new drugs as clinical trials are even more expensive than pre-clinical testing.

Survey of the quality of experimental design, statistical analysis and reporting of research using animals.

For scientific, ethical and economic reasons, experiments involving animals should be appropriately designed, correctly analysed and transparently reported. This increases the scientific validity of the results, and maximises the knowledge gained from each experiment. A minimum amount of relevant information must be included in scientific publications to ensure that the methods and results of a study can be reviewed, analysed and repeated.

Breeding and housing laboratory rats and mice in the same room does not affect the growth or reproduction of either species.

Few data exist regarding the effects of long-term housing of rats and mice in the same secondary enclosure. Historical reproductive and growth data were compared for colonies of mice and rats maintained in open-topped cages in either single-species or dual-species barrier rooms. This analysis included reproductive parameters (litter size at birth, litter size at weaning, and pups missing at weaning) collected from 33 colonies of mice comprising 500 to 38,500 breeding females and 28 colonies of rats totaling 350 to 4,600 breeding females, and representative samples from 28 colonies of each species were analyzed for weight gain from weaning to adulthood.

Quiet mutations in inbred strains of mice.

The year 2009 is the 100th anniversary of the founding of the first inbred strain of mouse, called DBA. During the last 100 years, inbred strains have proved their value for biomedical research and the number of such strains has mushroomed to over 450, each with different genotypic and phenotypic characteristics and useful for the study of disease and normal function. However, although inbred strains are stable, they are not fixed entities and researchers need to be aware of the phenomena of new mutations and of genetic drift, which occur within all mouse colonies.

Design and statistical methods in studies using animal models of development.

Experiments involving neonates should follow the same basic principles as most other experiments. They should be unbiased, be powerful, have a good range of applicability, not be excessively complex, and be statistically analyzable to show the range of uncertainty in the conclusions. However, investigation of growth and development in neonatal multiparous animals poses special problems associated with the choice of "experimental unit" and differences between litters: the "litter effect.

The origins and uses of mouse outbred stocks.

Outbred mouse stocks, often used in genetics, toxicology and pharmacology research, have been generated in rather haphazard ways. Understanding the characteristics of these stocks and their advantages and disadvantages is important for experimental design. In many studies these mice are used inappropriately, wasting animals' lives and resources on suboptimal experiments.

The choice of animal model and reduction.

Careful choice of the animal model is essential, if research is to be conducted efficiently, by using the minimum number of animals in order to provide the maximum amount of information. Inbred strains of rodents provide an excellent way of controlling and investigating genetic variation in characters of interest and in response to experimental treatments. Outbred stocks, in which genetic and non-genetic factors are inextricably mixed, are much less suitable, because random and uncontrolled genetic variation tends to obscure any treatment responses.

Is the use of animals in biomedical research still necessary in 2002? Unfortunately, "yes".

The use of laboratory animals in the year 2002 is essential both to maintain human health and to develop new treatments for the many diseases that still plague humans. The suggestion by Greek and Greek in Sacred Cows and Golden Geese in 2000, that animal experiments are invalid because animals are different from humans, shows clearly that they do not understand the philosophical basis for the use of models in science and every day life. Models only need to resemble the thing being modelled (the target) in a few key respects.

Refinement and reduction through the control of variation.

The key to doing animal experiments efficiently, while using the minimum number of animals without loss of scientific information, lies in good control of random variation, and recognition and control of "fixed effect" variation, such as the sex or strain of the animals. However, many scientists erroneously assume that the use of outbred, genetically heterogeneous animals is justified, because in some way, they more closely model humans. Unfortunately, all this does is to increase the phenotypic variation, which results in less-powerful experiments.

Good experimental design and statistics can save animals, but how can it be promoted?

Surveys of published papers show that there are many errors both in the design of the experiments and in the statistical analysis of the resulting data. This must result in a waste of animals and scientific resources, and it is surely unethical. Scientific quality might be improved, to some extent, by journal editors, but they are constrained by lack of statistical referees and inadequate statistical training of those referees that they do use.