Assessment of postsurgical pain in dogs (Proceedings)

It is now a settled matter that the adaptive capacities of animals, coupled with the innate biases of human observers, seriously impairs our ability to "know" which of our patients are in pain, how much they are in pain, and sometimes, even where they are in pain. Historically the absence of behaviors easily associated with pain (crying, whimpering, etc.) has been equated with the absence of pain. In fact, animals can lay quietly in pain, or conversely act painful until the approach of a person whereupon the painful behavior is replaced by a tail-wagging greeting.

Therefore it is first incumbent for veterinary clinicians to operate under the assumption that some procedures and conditions are inherently painful, without the patient having to "prove" they are in pain. Any surgical procedure, trauma, and many medical conditions such as gastroenteritis, pancreatitis, cystitis, inflammatory bowel disease, and certain neoplasms are all examples of where it is incumbent upon the clinician to include pain management in their treatment plan.

In fact, scoring pain is increasingly recognized as the "4th vital sign" in animals after temperature, pulse, and respiration (in humans it is the 5th , after blood pressure). Indeed the AAHA certification guidelines now require this assessment on every patient (MA23, PM1).

Overt signs of pain vary by species, and even within species some breeds and individuals have enhanced local and descending inhibitory mechanisms when compared others. Also, realize that pet owners may be better at "reading" their pet than a stranger. Listed below, are just some of the "new onset pain behaviors" shown by dogs in acute severe pain. (adapted from Karol Matthews excellent chapter in the Veterinary Clinics of North America Small Animal Management of Pain, July 2000 Vol 30 issue entitled: Pain Assessment and General Approach to Management). However, the observer must be very astute if they are to pick up signs of moderate or mild pain, and are exhibited not by the onset of new behaviors, but rather absence of usual behaviors.

An investigational tool in human medicine is evaluation of facial expressions, on the premise that certain muscle movements universally occur in our species when pain is present; this has been called "the primal face of pain." Computer software programs are being developed to detect these expressions that are then translated into a automated score. Similar primal facial expressions of pain are thought to exist in animals, e.g. furrowed brow, squinted eyes, ears turned back or away from the forward position, and even without a computer program these subtle changes can provide additional information to the veterinary observer.

Scoring pain in non-verbal patients is a special challenge, examples of which includes not only animals but neonates infant children and incapacitated (physically or mentally) adult humans. Measuring objective physiologic parameters has proven to be unreliable as indicators of pain. largely because of the influence of other non-pain influences e.g. stress, distress, anxiety, and normal biologic variation. Therefore we are left with subjective evaluations and such scoring systems should meet the following criteria:

     • interobserver variability and observer bias is minimized

     • they can distinguish varying levels of pain intensity

     • the degree of "importance" of pain to the subject is detected

Several terms that refer to qualities of a pain rating tool include: validity (the ability of the scale to effectively measure what it is supposed to measure), responsiveness (whether the tool can detect a change in pain, particularly one that is meaningful to the subject), and reliability (whether two observers will give similar ratings using the tool). A pain scale should also ideally be multidimensional, in that several aspects of the pain intensity or pain related disability are rated. Pain is felt to have multiple dimensions, and in human chronic pain rating tools, subjects may be asked a number of different questions about their pain - to rate the intensity, to describe how much it interferes with work or with family relationships, and to indicate how unpleasant the pain is. The importance of, or degree of unpleasantness of pain, or of any symptom may be related to the impact on survival that a given symptom has.

Despite the challenges, advances have been made in the use of behavioral evaluations to assign acute (especially post-operative) pain scores, falling under the following general categories:

SDS (Simple Descriptive Scale), where the observer picks value on a 0-4 or 0-10 scale (no pain to mild pain to moderate pain to severe pain to worst possible pain) that most accurately describes the patient.

VAS (Visual Analogue Scale), where a point is picked along a graduated 0-100mm "ruler", with 0 being no pain and 100 being the worst possible pain; classically employed, it can only be used by a patient that can self-report (i.e. is verbal, not animals, neonates, stroke victims, etc.), although sometimes an observer is used to pick the spot as a proxy, as would be necessary in veterinary medicine.

DIVAS, an expansion of the VAS which includes dynamic (the animal is aspected to move) and interactive (the wound is palpated) observations

NRS (Numerical Rating Scale), where values are assigned to several designated clinical criteria, the sum of which gives the pain score.

Once verbal descriptions of behavior become part of pain scales, then care must be taken to ensure that all potential users of the scale assign equal meaning and importance to the words or terms in the description. Arbitrary word meanings can increase the subjectivity of the measurement. With both VAS and NRS the observer's skill of evaluation is crucial to accurate assessment. Although an observation is recorded rapidly, VAS, NRS and SDS are considered to have poor reliability when multiple, less skilled observers are using them; advantages and disadvantages are commonly discussed. Many observers will never have seen an animal experiencing the "worst possible pain", and so the VAS scale is subject to bias. In addition, the values that are generated from these scales are non-numerical, thus it cannot necessarily be assumed that a subject whose pain was rated at 8 was twice a painful as one whose pain was rated at 4.

The advantages of the SDS and the VAS systems are inherently found in their simplicity and ease of use. However, they also historically result in the most variability, especially between observers. However, both inter- and intra-observer variability is diminished by adding in dynamic and interactive observation, that is, without and then with patient interaction, asking the patient to move, and palpation of the surgical site.

A pain scale that would take into account the various dimensions of pain would be theoretically more useful in indicating how much the pain "meant" to the animal, but VAS, NRS and SDS scales are said to be unidimensional. An alternative, multidimensional type of scale is the composite measurement scale (CMS), constructed such that it takes into account such dimensions as the temporal patterns, location, interference with basic function, or of enjoyment of life. These have been widely developed for human patients although these require the ability to answer questions on the impact pain is having on them. In veterinary medicine we would instead use a basic but essential understanding of pain rating tools and animal behavior, and combining those with actual "in the trenches" experience with observations. Such scales would ideally "ask" the subject to evaluate how much their pain meant by observing their willingness to do context specific things,. In theory, the CMS approach of looking at several indicators of well-being (e.g., posture, body weight or food intake, motor impairment, evoked pain responses and social interaction, for example) might also hedge against the problem of a single measure being confounded by unexpected factors.

The CMS considered most validated and in most widespread use in veterinary medicine is the Glasgow Composite Measure Pain Scale (CMPS). It identified and defined over 100 descriptive terms used by veterinarians to describe the behavior of dogs experiencing discomfort, and placed them into categories. These numerous descriptors were designed to minimize bias and interpretation by observer; it has been found to be most accurate for acute musculo-skeletal (trauma) and post-operative pain, rather than painful medical conditions. It has been published as a long (and rather time-consuming) form, and more recently as a short form which speeds patient evaluation considerably. Each of several categories of descriptors are then scored, and the combined value gives the final score. The maximum score is 24, and the closer to this maximum the more painful the patient.

There are several advantages that are common to the various scoring systems such as the Glasgow CMPS. One is that they force the observer to evaluate the patient in ways they may not ordinarily undertake. Another is that they recognize that animals will behave differently when they have human

Whatever system is used, it should include the following features 1) an observation of the patient without interacting, 2) an observation of the patient while interacting, and 3) palpation of the painful site. The author uses a hybrid DIVAS/SDS system for its simplicity, however inclusive of the 3 features above, and placing overall assessment of pain on a 0-10 scale.

The importance of scoring pain in animals may appear to be self-evident, but in human medicine at this time, no definitive improvement in outcomes can currently be attributed to pain scoring systems (which are becoming increasingly required as part of hospital and clinical certification programs). One study failed to associate improved Emergency Room analgesia for children with fractures and burns when mandatory pain scoring was implemented in triage, and another study concluded that pain scoring actually contributed to patient mortality (note: the author's conclusion was that the additional fatalities were a result of patient over-sedation for procedures, resulting from incorrectly elevated pain scores). In veterinary medicine, there exists no outcome measurements correlated to pain measurement and scoring. Nevertheless, as measurement modalities become more sophisticated, validated, and standardized, there is promise for outcome improvement. Furthermore, it is the author's strong viewpoint that if the only outcome of implementing pain scoring as the 4th vital sign is that it moves pain and pain control to a hospital staff's top of mind awareness, then it will have accomplished its goal. A 2004 UK survey revealed that 80% of the nursing staff agreed that pain scoring was a useful clinical tool (and usually scored patients more painful than veterinarians), but that only 8% of practices utilized a formal pain scoring system.