To be valid, methods used must assess the experience of animal pain and NOT some other experience.
'Gold standards' for measurement of an animal's experience of pain are NOT available and validation of particular indices or combination of indices may only be achieved by demonstrating a consistent relationship between changes in the indices and an established standard.
Two examples used for validating methods for assessment of animal pain under experimental conditions are highlighted below:
Formalin test (Inflammatory Pain model)
Abbott and others (1995) refined the pain rating scale for the formalin test (a model of inflammation used for testing analgesics, Dubuisson & Dennis 1977; Coderre and others 1993). The response has two phases, which reflect different parts of the inflammatory process (Abbott et al 1995). They have continued to optimise the formalin dose and conditions under which the test is conducted to improve animal welfare, by reducing the variability within the test and therefore the number of animals required, and the severity of pain (Abbott et al 1986, 1999).
Increasing concentrations of a formalin solution were injected into the rear paw and the behaviour was recorded for 60 minutes. There was a positive correlation between the concentration of the formalin and the time spent licking, lifting and favouring the injected paw. The time spent in different paw directed behaviour also changed with the concentration e.g. at low formalin concentrations favouring the paw was the dominant behaviour but as the concentration increased this behaviour was replaced first with paw lifting then paw licking. The sum of the times spent lifting and licking the paw was superior to any single measure. General grooming and paw flicking were also increased with increased formalin concentration, while general activity and exploring were decreased.
The test was further validated by showing that analgesic treatment (known to be effective in rats) consistently reduced the response and the sum (index) of the time spent favouring, lifting and licking the paw.
Molony et al (2002) decreased the amount of tissue (from routine castration and tail docking of lambs less than one week of age) made hypoxic and ischaemic by rubber (elastrator) rings. The authors showed that decreasing the amount of ischaemic visceral tissue (from 2 testes, 1 testis, no testes) decreased the amount of lying with full extension of the hind limbs, and increased the time spent in normal lying and standing postures.
The incidence of active behaviours also decreased as the amount of ischaemic tissue was reduced or the peripheral tissue blocked with local anaesthetic.
For quick, accurate, and efficient assessment of animal pain, detection of only two levels of pain may be sufficient (no pain, painful requiring treatment). If you are establishing your own measuring scheme determine the validity, reliability (every time) and sensitivity (will the index detect mild as well as severe pain) of the indices in the scheme before using it in practice. You may validate your indices by making measurements after a painful procedure carried out with and without analgesia, anaesthesia, known to be effective for the type of pain being studied in the species of interest, as long as you do NOT then use the indices to assess the effectiveness of analgesics.