Comments & suggestions  should be directed to Visit my website SELECTING ANIMAL MODELS The use of animals in research and instruction generally occurs in one of two contexts: the animals serve as model systems for the investigation of phenomena and processes which cannot be studied directly, or the animals are being studied to investigate a problem specific to the particular species.   Most biomedical research falls into the first category. Examples of the latter include field studies of behavioral and ecological adaptations of animal species, studies of taxonomic relationships among species, or captive studies of physiological or behavioral processes which form an important part of the adaptations of one or more species. This section reviews scientific, ethical, and humane considerations in experimental design and model selection for research involving animals.  Ethical and humane considerations should be viewed as compatible with good scientific practice. There is a body of literature that supports the premise that animals which are humanely cared for are healthier, both physically and psychologically, and therefore make better, more predictable, subjects.  Similarly, unless a research project is intended to study pain itself, pain and suffering on the part of the animal subject will rarely, if ever, contribute anything positive to the experimental procedure.  Thus, ethics and humane considerations can be viewed as integral parts of the process of experimental design and model selection.  Selecting a Research Model Selection of an appropriate model must be based on extensive familiarity with the problem or system to be studied, so that the researcher can determine the range of biological responses necessary to the experimental design.  Once this familiarity is developed, either by extensive review of the published literature or from pilot studies, the researcher can proceed to select an appropriate model, whether a whole animal, animal-derived material or a non-animal model. Types of Models Whole animal models are usually chosen when the system being studied can best be understood in the context of its interactions with other systems in the organism (e.g., sexual differentiation in embryonic development) or when it is the organism as a whole which is the system to be studied (e.g. the ontogeny of aggressive behavior).  Some systems are better studied in isolation in animal cells, tissue or organs.  For example, a number of biochemical and cellular processes can be studied in tissue or organ cultures derived from animal material.  For other kinds of studies, statistical or computer models may be appropriate.  It should be obvious, however, that the data generated from such non-animal models are only as good as the data upon which the models are based; thus, animal studies of some kind are prerequisite for developing and verifying all models. Factors Affecting Choice of Species An animal model is a living organism in which normal biological processes can be studied, or in which a spontaneous or induced pathological process can be investigated.  To be effective, the process being modeled should closely resemble the analogous process in human beings or some other species in one or more ways.  Some important criteria of animal models are: relevance to the problem being studied; the accuracy with which the model reflects all or some important aspects of the problem; the model's predictability; and the model's availability to researchers.  In addition, general species characteristics such as life history parameters, behavior and diet can be as important as physiological parameters in species choice. Life History Aspects of life history patterns which may influence species choice include developmental rate, age at first reproduction, frequency and timing of reproduction, gestation length, litter size and life span. Species with relatively short life spans are sometimes the most useful choice for models of the long term effects of early ontogenetic factors.  Life history parameters may also combine with other species characteristics to influence the appropriateness of model choice.  For example, litter size is an important consideration when studying a naturally occurring genetic condition as a large litter will increase the likelihood that some offspring will carry the trait. Behavior The normal behavioral patterns of a species can be important to model choice, whether or not the researcher is actually interested in the animal's behavior.  The normal social organization of a species affects such variables as how animals must be housed or fed and under what circumstances they will reproduce successfully.  Social and individual behavioral characteristics may influence research variables in both obvious and subtle ways.  An example of an obvious behavioral constraint is that, with many species, placing more than one adult male in a social group under the relatively confined conditions of captivity will result in serious aggression that may disrupt research.  A more subtle effect has been observed in mice where females, but not males, show enhanced immune responses when housed in groups.  Such behavioral effects may introduce uncontrolled variables into experiments if they are not anticipated by the researcher. Diet Knowledge of the normal nutrient requirements of a species is an important factor in selection of an appropriate model. Use of poorly nourished animal subjects, or subjects which lack some critical dietary component, may introduce extraneous, uncontrolled variability into experiments. Genetics (top of page) Genetic factors are important in model selection in several respects.  Species selected should have a well known backgrounds, as some may develop naturally occurring genetic diseases.  In some common laboratory species, mutant or inbred strains have been developed with well documented and often highly specific genetic properties. Alternatives to the Use of Animals in Research Critics of animal research have suggested that most, if not all, uses of animals in research and education could be eliminated by the use of alternatives, such as tissue culture or computer models.  In 1986, the U.S. Congress Office of Technology Assessment produced an extensive study of the use of animal models and options for alternatives to animal use in research, education and testing.  The general conclusion of the report was that the very nature of the research in many areas makes it highly unlikely that reasonable alternatives to animal use will be developed.  A working definition of alternative techniques is “those techniques which replace the actual use of animals, reduce the numbers used, and/or refine the techniques to minimize the potential for the animal to experience pain or distress.” Congress has been urged to provide funds specifically to develop alternatives to animal use in research, testing and education. Legislation has included requirements that alternatives be considered (e.g. the Health Research Extension Act of 1985 and the amended Animal Welfare Act) and NIH has a funding program for the development of methods which replace, reduce, or refine animal use. Replacement Alternatives which replace animal models can be classified into the following broad general categories: Modification of Existing Animal Use It is sometimes possible to substitute one species of animal for another.  For example, lower vertebrates or invertebrates may be substituted for higher vertebrates, or "laboratory species" (such as mice or rats) may be substituted for animals such as dogs or cats.  Such substitutions are usually advocated on the grounds that species differ in their capacity to suffer pain or distress and it is assumed that invertebrates or lower vertebrates will suffer less than higher vertebrates. If the research is to yield useful results, the animal species selected must be one which fulfills the requirements of the model and closely mimics the condition being studied.  The relative capacity of different species to suffer pain or distress is intensely debated by animal welfare advocates and scientists.  Selection of a particular species for a research project is constrained by many considerations and there may be few or no other species which satisfy all requirements of the model.  In biomedical research, where  models are often of human diseases, invertebrates or lower vertebrates may share so few relevant characteristics with humans that substitution is impossible.  In basic biological research, the problem being studied may be specific to a particular species or group of species and may not occur in other groups. Plants and microorganisms have been suggested as substitutes for animals.  For example, Salmonella is used in mechanistic studies in genetics, and the active steroid hormones found in yeasts are used in some endocrinological and immunological research.  Use of plants and microorganisms as substitutes is limited by their evolutionary differences from humans and other higher vertebrates, and by their own unique characteristics. Improvement in experimental design, or in  statistical analyses of results, may reduce the number of animals needed.  Knowledge of statistics, can improve experimental design by including consideration of such factors as randomization, confounding variables, sample size and statistical power.  Animals are expensive to use, so experimenters usually employ the minimum number of subjects, raising the possibility that improved design or statistics may actually increase, rather than decrease, the number of animals used.  If too few animals are used, experimental results may be statistically invalid and therefore useless.  Improved design and analysis could reduce animal use if more robust or clearer results reduced the total number of experiments, or eliminated unnecessary duplication of experiments.  Reproducibility of results is a key component of the scientific method.  Therefore, all repetition of experiments cannot, and should not, be eliminated. In some cases, experimental design can be modified to reduce or eliminate pain and distress of the animal subject.  For example, noninvasive systems for sampling physiological parameters are sometimes used to replace invasive methods.  In behavioral research allowing a subject increased control over painful or unpleasant stimuli may reduce distress for some species. Use of Animal Derived Material Although critics of animal research see alternatives as a way to eliminate animal use in research, many suggested "replacements" consist of animal derived material.  Examples of the use of animal derived material include cell, tissue and organ cultures.  Working with culture specimens avoids potentially painful manipulations of live animals, although these materials must originate in a living animal.  Use of such models may reduce the number of animals needed for research.  In some cases cell culture studies can utilize cell lines derived from propagation of relatively few cells.  Ultimately, however, hypotheses and data derived from these must be checked in the whole organism. Chemical and Physical Models In some instances it is possible to use physical or chemical models to study living systems.  The study of many biochemical mechanisms, for example, make use of materials isolated from organs or tissues. Some physical and mechanical models have been developed mainly for educational and training uses. Mathematical and Computer Models (top of page) Whenever a function or a relationship within a living system can be described mathematically, the possibility exists for developing a mathematical model.  Scientists have long employed such models in biological and medical research because they provide the opportunity to vary the parameters involved and to predict what effects different parameters will have on the system.  However, the comprehensive  knowledge of parameters required by such models is not available until after extensive experimental work has been done on a living system.  Furthermore, the final stage of this process must be to go back to the organism to assess the accuracy of the model's predictions. It has been suggested that computer technology is now so advanced that computer based models can completely substitute for the use of animals in research.  Computers can be used to analyze data but they do not generate knowledge.  To make use of a computer model researchers or instructors must supply the computer with whatever information is needed for the model.  If a living system is being modeled, the only source of this information is the living system itself.  The more detailed the information supplied to the model, the better the model is likely to be.  Based on information derived from animal studies, computer models have been developed to analyze relationships within and between living systems.  Computer models have been particularly useful in modeling feedback systems.  Biomedical applications of computer models include aspects of kidney, cardiac and lung function, regulatory systems, endocrine function, sensory physiology, neurophysiology and developmental biology.  In animal behavior, for example, game theory has been used to construct computer models which would predict how animals might behave during aggressive encounters. It should be noted that many of these alternatives are not new.  The history of the use of animals in research can be viewed as an ongoing process of the refinement of use of animal models with many factors (ethical, humane, scientific and economic) driving the refinement process. Reduction In discussing the ways to reduce the numbers of animals used, the definition of an animal and the principle of moving down the phylogenetic scale must also be kept in mind. Three broad categories for reducing the number of animals used include: Animal Sharing Sharing of animals can significantly reduce the number of animals used within a given institution.  Sharing can be as simple as allowing someone to practice a surgical approach on an animal that has been, or is to be euthanized for other purposes, or providing organs or tissues at the time of necropsy.  Sharing becomes more complicated when attempting to maximize the use of control animals, but it can significantly reduce the number used within an institution.  If two studies involve the need to perform the administration of compounds by identical routes, the use of standard control diets, or the need to condition animals to a particular environment, control animals could be shared  within the institution. Improved Statistical Design Anyone who has ever taken a course in experimental design or applied statistics has been bombarded with the importance of consulting with a statistician during the design phase of the experiment and not when the data collected needs to be analyzed.  Improper design of experimental protocols and/or the failure to use appropriate statistical methods can result in the usage of an inappropriate number of experimental animals.  A variety of design strategies are available which can reduce the number of animals needed in a given study. Experimental protocols which utilize serial sacrifice, group sequential testing and crossover designs can significantly reduce the numbers of animals required.  The availability of low cost statistical packages  permits investigators access to sophisticated data  management and analysis.  This accessibility makes possible the use of design criteria and complicated statistical analysis which previously have been confined to institutions with large statistical support units. With this ability at their finger tips, investigators should be able to maximize the analysis of the data generated from each animal used, thus reducing the total numbers of animals necessary for a particular set of data. Phylogenetic Reduction (top of page) Projects which can be designed to use invertebrate species instead of vertebrate species represent a type of phylogenetic reduction which was discussed as a replacement technique.  Such broad jumps across the phylogenetic scale are not always possible, but less dramatic shifts can significantly reduce the numbers of higher species being used in research, teaching and testing.  In many instances, the theory of phylogenetic reduction has been blurred by a species's use as a companion animal with little regard for phylogenetic ranking.  The animals chosen for project usage should be the least advanced from a phylogenetic standpoint that will provide the necessary data.  The principle of phylogenetic reduction is generally well accepted as a way to reduce the number of animals used, but it often brings many hidden difficulties.  As one descends the phylogenetic scale, the available information on the maintenance and use of these animals often becomes difficult, if not impossible, to obtain.  When choosing a study model, it is critical that the principal investigator take into account their ability to provide appropriate care for the species chosen.  Phylogenetic reduction is an important means of decreasing the number of animals used, but should be practiced carefully and with the full knowledge of the requirements of the species chosen. Refinement Refinement refers to techniques which reduce the pain and distress to which an animal is subjected.  For the purpose of this manual these techniques can be classified into the following broad categories: Decreased Invasiveness A hallmark of most of the new diagnostic and therapeutic techniques used in human medicine is the minimal degree of invasiveness that is required to successfully perform a procedure to obtain a given set of data.  In many instances these techniques are applicable in the research environment and can be adopted for use in animals.  A sophisticated example could be the use of Magnetic Resonance Imaging for results that formerly required euthanasia of multiple animals along a time curve to obtain assay tissue. Today one animal can provide all the information along a given curve.  A less dramatic example is the vascular access device which permits repeated samples or injections in a single animal instead of using several animals.  Invasiveness reduction methods are available in almost every area of biomedical research, and in project design, it is important to identify and use these methods wherever possible.  Not only do they represent an alternative technique, but they generally provide much more consistent and reproducible data. Improved Instrumentation In this age of microelectronics, fiber optics and laser instrumentation, the potential for refining techniques used in animal experimentation seems almost limitless.  Improved instrumentation can minimize animal distress by reducing the level of restraint and/or manipulation necessary to obtain biological samples.  Included in this category are the use of tethers in a variety of species to allow continuous access to the various organ systems, while permitting the animal virtually unrestricted movement within its primary enclosure.  The advantages of these systems are numerous, not the least of which is minimizing a variety of nonexperimental variables associated with prolonged restraint. Once obtained, samples can be analyzed in very small volumes for a multitude of parameters.  Examples of this can be found in the commercially available diagnostic laboratory equipment which require only microliter blood samples to perform a variety of diagnostic tests.  The use of smaller sample sizes permits the use of smaller animal species and prevents the need to euthanize many of these species to obtain the necessary volume of blood.  It is now possible to obtain serial blood samples from small laboratory rodents which reduces the number of animals necessary to obtain data over the length of the study. Improved Control of Pain The Animal Welfare Act requires "that the principal investigator consider alternatives to any procedure likely to produce pain or distress in an experimental animal" and in any practice which could cause pain to animals that a doctor of veterinary medicine is consulted in the planning of such procedures for the use of tranquilizers, analgesics and anesthetics. Improved Control of Techniques (top of page) Proficiency in the handling and restraint of animals makes it easier to perform a variety of routine procedures with minimal or no pain or distress to the animals involved. Animals are creatures of habit and when proper handling is part of their regular routine, the degree of distress caused by the procedures is minimized.  Animals can be trained or conditioned to accept a variety of procedures which if suddenly forced upon them can be distressful.  Almost every animal commonly used in the laboratory responds positively to humane treatment.  To develop the proper techniques and gain confidence in their use requires training by someone with appropriate experience.  This can be the veterinarian, a member of the animal care staff or a fellow investigator.