Effective pain management can only be achieved and maintained when signs of pain can be assessed reliably and accurately. The experience of pain is unique to each individual, and in man this experience can be described verbally to facilitate pain management. However in animals pain assessment is difficult, often resulting in:

• Analgesics being withheld because the animal is not showing overt signs of pain

• An assumption that since a human who has undergone a similar procedure would require analgesics an animal must also require pain relief, and therefore analgesics are administered

The result of the first scenario is animals that are almost certainly in pain will not receive analgesia. In the second case, since no proper assessment of the degree of pain is made, it is a matter of chance whether an appropriate degree of pain relief is given. In addition, although an initial dose of analgesic is usually beneficial, and unlikely to be detrimental to the animal, in the longer term repeated inappropriate analgesic administration may also cause side effects. Therefore it is clearly preferable to try and assess the animal's pain and adjust the analgesic regimen on the basis of the assessment.

Numerous different methods of assessing pain have been developed in man and attempts have been made to apply some of these to animals. In verbal humans, scoring systems or pain questionnaires (e.g. McGill pain questionnaire) are used to evaluate pain and manage analgesic administration directly. However in infants and babies, non verbal or written communication is not possible, leading to similar difficulties with pain assessment to those found in animals.

It is known that objective measures, including heart rate, respiratory rate and body temperature are an unreliable guide to the presence of pain (Holton et al.1998). Therefore current methods to assess clinical pain in animals are based primarily on behavioural observations. However many factors can also influence behaviour adding to difficulties in interpreting how much pain an animal is experiencing. Animals undergoing diagnosis and treatment of clinical disease are usually in an altered environment, leading to changes in their normal behaviour that may mask signs of pain. It is important to interpret pain in each species based on the normal behaviour of that species. This is particularly noticeable when pain behaviour in dogs and cats is compared. Cats rarely vocalise in response to pain, but can become depressed and sit hunched at the back of the cage, leading to the false assumption that the animal is not in pain. The attention seeking behaviour of many dogs in pain leads more readily to the administration of analgesic drugs.

Recognition of pain in horses is also difficult and presents a challenge to anaesthetists and equine practitioners in determining analgesic drug administration. Few studies have investigated postoperative pain in horses. Johnson et al. (1993) compared the effects of three NSAIDs after elective orthopaedic surgery, using a score, awarded after subjective evaluation of behaviour by one of two assessors. McCarthy et al. (1993) measured plasma b-endorphin concentration to assess pain and stress, but considered it too non-specific to be useful. Raekallio et al. (1997a) also attempted to quantify perioperative pain in horses after elective orthopaedic procedures. Pain was scored subjectively, and by measurement of predefined behavioural and physiological values. Later Raekallio et al. (1997b), using the same scoring system, investigated the analgesic efficacy of phenylbutazone in horses after arthroscopy. A poor correlation between the variables examined and perioperative pain was found (Raekallio et al. 1997a, 1997b), highlighting the difficulties of pain assessment in this species.