Education Section

Pain measurement in humans

K.S. Ong and R.A. Seymour*
National University of Singapore, Faculty of Dentistry, Department of Oral and Maxillofacial Surgery, Singapore. *University of Newcastle-upon-Tyne, Faculty of Dentistry, Department of Restorative Dentistry, Newcastle, U.K.

Correspondence to: K.S. Ong, 435 Orchard Road, Suite 11-02, Wisma Atria, Singapore 238877 Email: cliffong@pacific.net.sg

Introduction Basic concepts Pain measurement instruments

Multidimensional pain scales Factors Recommendations References

Keywords: Pain assessment, pain scores, pain scales, acute pain, chronic pain
Surg J R Coll Surg Edinb Irel., 2 February 2004, 15-27

Sound measurement, an essential component of any scientific discipline, remains a particular problem in pain research. The measurement of pain intensity, for example, is a difficult and often a subjective undertaking. This is of little surprise to clinicians and researchers, because it is well recognised that pain intensity, like other sensations and perceptions, is a private experience that displays considerable variability both across patients and within a patient across time. Nonetheless, pain measurement and discerning factors that may affect its measurement are important for diagnosis and to determine the effectiveness of treatment interventions. This article reviews the basic concepts, roles, instruments used, and factors affecting pain measurement. A variety of the most commonly used pain measurement instruments are evaluated for their advantages and disadvantages. The article aims to assist clinicians and researchers to select the pain measurement instruments that best serve their purposes

INTRODUCTION
The definition of pain, according to the International Association for the Study of Pain (IASP), is “an unpleasant sensory and emotional experience associated with actual or potential tissue damage and described in terms of such damage.”¹ A critical aspect of the IASP’s definition is that pain is defined in terms of human experience. If the study of pain in people is to have a scientific foundation, it is essential to measure that human experience. If we want to know how effective a new drug is, we need numbers to say that the pain is decreased by some amount. In attempting to measure pain, the clinician must first decide from which viewpoint he or she will observe the pain. To measure pain in humans, it is essential to rely on language communication. A common language that can be used in a consistent manner allows different clinicians and researchers to communicate.

The literature on this topic is extensive, and a complete review is beyond the scope of this article. Instead, the authors seek to provide an introduction to the current literature of pain measurement by highlighting selected topics and studies. The topics are chosen to give the readers a general picture of the controversies in the area of pain measurement. Our aim is to assist clinicians and researchers to select the pain measurement instruments that best serve their purposes.

The topics will be discussed as follows:

• Basic concepts in pain measurement

• Role of pain measurement in clinical and research activity

• Ideal pain measurement system

• Pain measurement instruments

• Factors affecting pain measurement

• Recommendations to good clinical pain measurement.

BASIC CONCEPTS IN PAIN MEASUREMENT
A critical aspect of the IASP’s definition is that pain is defined in terms of human experience, and its presence can only be communicated through the linguistic descriptions of this experience. There appears to be at least three distinct dimensions of the pain experience that can be assessed: sensory-discriminative, affective-motivational and cognitiveevaluative.² The sensory-discriminative dimension comprises the sensory aspect of pain, including intensity, location, and temporal aspects. The affective-motivational dimension reflects the emotional and aversive aspects of pain and suffering. The cognitive-evaluative dimension reflects the patient’s evaluation of the meaning and possible consequences of the pain and injury, including impact on quality of life and even death itself.

According to an accepted model, pain consists of four components, namely nociception, sensation, suffering and behaviour.³ When studying clinical pain, it is often impossible to examine the relationship between nociception and the pain response, since in many cases there is no pain stimulus. Consequently, to assess the pain experience, measurement instruments have traditionally relied upon subjective reporting, focusing on sensation, suffering and behaviour rather than on nociception. Since pain is a personal experience, the only way the clinician can judge a patient to be in pain is: (1) to rely on the patient’s self-report, (2) to observe resultant behaviour, and (3) to measure the physiologic parameters that we believe to be characteristic of a patient in pain.

Acute pain differs from chronic pain, not just in duration and intensity or the fact that acute pain is more closely linked to identifiable pathology, but because the noxious stimulus is more readily identified. In chronic pain, particularly, there is no correlation between the amount of pathology and the degree of pain. Pain behaviour lingers, often long after the pathologic process has healed. These differences also mean that the pain measurement required in these two types of pain is different.

In general, a pain measure test should be reasonably simple to administer and directed at a level that most patients can understand. It should be accurate and reliable. Also, it should fulfil the criteria of validity and sensitivity and possess reliable scaling properties. Of these characteristics, reliability and validity are the two most important.⁴ An instrument is reliable if the test scores provided by the same individual on two separate occasions are similar. Validity is the extent to which a test measures what it is supposed to measure. A scale should be as sensitive as feasible. If intensity is the particular quality of pain that is to be measured, then the scale should cover the whole dimension of intensity.

ROLE OF PAIN MEASUREMENT IN CLINICAL AND RESEARCH ACTIVITY
Pain measurement provides valuable clues that help in the differential diagnosis of the underlying causes of pain.⁵ Description of the burning qualities of pain after peripheral nerve injury or the stabbing, cramping qualities of visceral pain frequently provide the key to diagnosis. Many of the test instruments have discriminatory merit and may aid the clinician in diagnosing a specific condition. For example, the McGill Pain Questionnaire (MPQ) can serve as an aid in the differential diagnosis between various pain syndromes. A study was carried out to demonstrate the discriminative capacity of the MPQ, which was administered to 95 patients suffering from one of the eight known pain syndromes: post-herpetic neuralgia, phantom limb pain, metastatic carcinoma, toothache, degenerative disc disease, rheumatoid arthritis, labour pain and menstrual pain.6 A multiple group discriminate analysis revealed that each type of pain is characterised by a distinctive constellation of verbal descriptors. Furthermore, when the descriptor set for each patient was classified into one of the eight diagnostic categories, a correct classification was made in 77% of the cases.

Pain measurement also helps to determine the most effective treatment, such as the types of analgesic drugs or other therapies necessary to control pain, and are essential in evaluating the relative effectiveness of different therapies.⁷ If we want to know how effective a pain treatment modality is, we need numbers to say that the pain is decreased by some amount. The initial measurement of the pain experience also produces a baseline on which to assess future therapeutic interventions. 

Pain measurement is important in pain research.⁸ It is important to decide if a new pain treatment technique is superior to another. The numbers obtained from the research can be subjected to statistical analysis to decide if there is a “statistically significant difference” between the treatment techniques. However, for research purposes, especially to distinguish the analgesic effect between two drugs, a finer grade or more sensitive painmeasuring instrument may be required to be able to detect the difference.

Pain measurement is also important in the assessment of the degree of disability or impairment of function related to pain, for therapeutic or compensation reasons.⁹ Impairment is a medical concept, and disability is a legal concept. Impairment is any lost or abnormality of psychological, physiologic, anatomic structure or function. Legally, disability refers to the inability of an individual to meet social or occupational demands because of physical or psychological disadvantage, or it refers to the fulfillment of statutory or regulatory requirements for compensation.¹⁰ This definition corresponds with the World Health Organisation classification of handicap.

IDEAL PAIN MEASUREMENT SYSTEM
Five properties of an ideal pain measurement system that have theoretical and practical advantages have been proposed by Gracely and Dubner (1981).¹¹ They are are outlined below:

• Sensitive measurement free of biases inherent in different assessment methods. The system needs to provide scale free metrics, which allow measurement from differing scales to be unified statistically. With the appropriate measurement system in place, it also enables comparisons of results across dissimilar measurement scales

• Provision of immediate information about the accuracy and reliability of the subject’s performance in the task. The system must be simple and fast and at the same time generate statistics that address the degree of “fit” not only for mean responses but more importantly for each separate response to a hypothetical model

• Separation of the sensorydiscriminative aspects of the pain experience from its hedonic (affective) qualities. This requires separation of inherent stimulus painfulness from the ratings an individual makes of that stimulus

• Usefulness for clinical as well as experimental pain measurement, allowing reliable comparisons between these fundamentally different types of pain. This can be difficult, one which may not reflect legitimate differences between researcher’s needs for precision and clinician’s needs for speed and global assessment

• Absolute measures that increase the validity of pain comparisons between groups and within groups over time. It must be able to generate appropriate tests of significance.

Currently, there is no single pain measurement instrument that fully satisfies all these properties. With the above properties in mind, the various pain measurement instruments available will be evaluated.

PAIN MEASUREMENT INSTRUMENTS
There is no simple thermometer that can objectively record how much pain an individual experiences. Anything that can be determined about an individual’s pain is based on what the patient verbally or nonverbally communicates. Assessment of the pain experience is frequently built upon the use of patient’s self-report. Because pain is subjective, the patient’s self report provides the most valid measure of the experience and is considered the gold standard in pain measurement.^1,12  The self-report relies heavily on the ability of the patient to recall pain. These are no more veridical than the patient’s own descriptive choice of language; if the patient was lying we could not detect this. Studies have shown, however, that patients do retain accurate recall of specific pain experiences.¹³

There are a large number of pain measurement instruments (perhaps hundreds) that have been developed for use. Only the more commonly used and tested instruments will be discussed in this review. Figure 1 shows some of the most commonly used instruments. These instruments can be categorised as shown in Table 1.

RAPIDLY ADMINISTERED UNIDIMENSIONAL PAIN SCALES

Figure 1: The most commonly used single dimension pain scales

The strengths of VRSs include the ease with which they can be administered and scored. Because they are generally easy to understand, compliance rates for VRSs are as good or better than those for other measures of pain intensity under most conditions. ¹⁶ Also, VRSs have consistently demonstrated their validity as indicants of pain intensity. They are related positively to other measures of pain intensity.^16,17 This scale has also claimed to be superior to other scales in assessing the effects of analgesics on acute pain.¹⁸

A limitation of this rank scoring method is that it assumes equal intervals between the adjectives, even though it is unlikely that equal intervals exist. That is, the interval between no pain and mild pain may be much smaller than that between moderate pain and severe pain, yet each interval is scored as if the difference were the same. A second problem is that VRSs represent ordinal scaling data. Verbal rating scale scores are often treated as if they were interval or ratio data, and subsequently analysed using parametric statistical procedures rather than the appropriate nonparametric procedures. A third problem is that it lacks sensitivity and does not allow for finer grade pain assessments. The small number of descriptors may force the patient to choose a particular category that may not describe the pain satisfactorily.

Numerical rating scale
Numerical rating scales NRSs involves asking the patients to rate their pain from 0 to 10 (11 point scale) or 0 to 100 (101 point scale), with the understanding that 0 represents one end of the pain intensity continuum (no pain) and 10 or 100 represents the other extreme of pain intensity (unbearable pain). The number that the patient states represents his/her pain intensity score.

The validity of the NRS has been well documented. The scale demonstrates positive and significant correlation with other measures of pain intensity.^16,19 They have also demonstrated sensitivity to treatments that are expected to impact upon pain intensity.^14,20 A recent study demonstrated that reduction of approximately two points (approximately 30%) on the NRS represented a clinically significant difference in pain reduction.²¹ This was done by relating the NRS to global assessments of change in multiple studies of chronic pain. Numerical rating scales are likewise extremely easy to administer and score, so they can be used with a greater variety of patients, e.g. geriatric patients, than is possible with the visual analogue scale. The simplicity may be the reasons for the high rate of comparative compliance with the measurement task.

The primary weakness of the NRS is the lack of research comparing their sensitivity to that of other measures, particularly the VAS.¹³ One study found that an 11 point NRS was less sensitive to the effects of cognitive-behavioural treatment than the visual analogue scale.²²

Visual analogue scale
The visual analogue scale (VAS) consists of a line, usually 10 cm long, whose ends are labelled as the extremes of pain (e.g., no pain to unbearable pain). A VAS may also have specific points along the line that are labelled with numbers or intensity-denoting adjectives. Such scales are called graphic-rating scales. Patients are asked to indicate which point along the line best represents their pain intensity. The distance from the no pain end to the mark made by the patient is that patient’s pain intensity score.

A VAS for pain affect has been developed in an effort to include domains of measurable pain experience other than the sensory intensity dimension. The patient is asked to rate the unpleasantness of their pain experience (i.e., how disturbing it is). Endpoints are labelled “not bad at all” and “the most unpleasant feeling imaginable”.

The VAS has the advantage of simplicity. It is widely used and is independent of language. It is easily understood by most patients and can be readily reproduced on successive presentations. Children from age seven can understand it.²³ It is sensitive to pharmacological and nonpharmacological procedures which alter the experience of pain.^14,24 It also correlates highly with pain measured on the VRS and NRS.^16,25 A major advantage of the VAS is its ratio scale properties and so may be treated as such statistically.²⁶ In contrast to many other pain measurement tools, equality of ratios is implied, making it appropriate to speak meaningfully about percentage differences between VAS measurements obtained either at multiple points in time, or from independent samples of subjects. The scale also has a high number of response categories. This makes the VAS potentially more sensitive to changes in pain intensity than measures with limited numbers of response categories. Research indicates that the VAS of pain intensity is usually more sensitive to treatment changes than VRS.^14,24,25,27

The addition of descriptions along the length of the line may affect the distribution of results.²⁸ Along a vertical scale, they will cause the majority of the results to be grouped around the descriptions. For those results, the scale behaves like a simple descriptive scale, and the sensitivity will be reduced.Numbers should not be placed on the VAS, since preferred numbers like 5 and 10 will attract an unfair share of the results. A plain line VAS (absolute or unmarked line) running from left to right is the most unbiased scale and is recommended.^28,29

Although, the VAS is easy to use, scoring can be slightly more difficult than VRS and NRS. Some patients may have difficulty understanding and using the VAS measures. There is a quoted failure rate of seven per cent.³⁰ Responses to VAS are influenced by several various biases affecting psychophysical responses. It requires a certain amount of visual and motor coordination, which may be lacking in the post-operative period, and measurement may be difficult to perform after anaesthesia, when patients may have difficulty concentrating.

Psychophysical measurement (descriptor differential scale)
To bridge the distance between laboratory and clinical approaches to pain measurement, cross-modality matching methods have been employed. In this method, a sensory experience, represented by a set of words that describe pain is quantified by matching a stimulus in another sensory modality, such as loudness of sound, or handgrip force. The concept of proportionality is central to the performance of the cross-modality matching procedure. It has been suggested that this procedure is likely to produce less bias and more consistent results than category scales measures.³¹

The descriptor differential scale (DDS) was developed hoping to remedy a number of deficiencies associated with existing pain measurement instruments. The DDS was designed to reduce bias, assess separately the sensory intensity and unpleasantness dimension of pain, and provide quantification by ratio scaling.

The DDS consists of two forms that measure separately the sensory intensity and unpleasantness qualities of pain. Each form consists of 12 verbal descriptors, in which each descriptor is centred over a 21-point scale with a minus sign at the low end and a plus sign at the high end. The patients rate the magnitude of the sensory intensity or unpleasantness of the pain they are experiencing. The reliability and objectivity of this instrument was confirmed by presenting it to a group of 91 patients undergoing third molar extractions.³² Sensory intensity and unpleasantness was administered to all patients one hour and two hours after surgery. Total scores on both forms showed high test-retest reliability coefficients as did scores derived from individual items. Correlation coefficients between individual items and the total score revealed a high degree of internal consistency for both forms of DDS.

Behavioural measure scales
Behavioural measure scales are especially useful for measuring pain in infants and preverbal children who lack language skills, adults who have a poor command of language or when mental clouding and confusion limit the patient’s ability to communicate meaningfully. Under these circumstances, behavioural measures provide important information that is otherwise unavailable from patient’s self-report. Most behavioural pain measures have been developed to obtain concise records of how children respond when they experience pain. Usually, a trained observer monitors children in pain and documents any behaviour that suggests discomfort such as crying, flailing, grimacing, and protectively guarding certain body parts. Itemised behavioural checklists/ scales are then developed listing the behaviours specific for each type of pain. Clinicians complete the checklist by noting which behaviours occur and how long they last. The rationale is that an objective evaluation of children’s pain behaviours should provide an accurate estimate of the strength of their pain experiences. A child’s pain level is calculated either by adding the number of observed behaviours or by first assigning a weighted distress value to each behaviour and then summing the individual values to produce a composite score.³³ The behavioural pain measures currently available for infants/children are listed in Table 2. Most of these behavioural measures are designed primarily to assess acute procedurebased pain. For example, the infant pain behaviour rating scale is a timesampling overt distress scale designed for measuring the infant’s vocalisations, expressive body movements and facial expressions during injections.³⁴

Picture scales
The picture scale employs eight line drawings that illustrate facial expressions of persons experiencing different levels of pain intensity.³⁵ Patients are asked to indicate which one of the eight expressions best represents his or her pain experience.

Each picture has a number (from 0 to 7) representing the rank order of pain illustrated, and the number associated with the picture chosen by the patient represents that individual’s pain score. The picture scale has yielded some, but as yet, limited, empirical evidence for its validity.³⁵ Because of the difficulty of assessing pain in young children, it is often necessary to rely upon behavioural parameters such as facial activity. One such method is the facial action coding system, in which facial activity is broken into 44 separate action units. The modified version employs a limited number of rated facial expressions believed to be consistent with pain expression.³⁶

Pain diary
A pain diary is a personal, self-reported written account of patient’s daily pain experience and behaviour. The patient is asked to make notes of pain intensity, particularly in relation to particular behaviour such as: daily activities for example sitting, standing, and lying down, sleep patterns, and pain medication taken. The information elicited from a pain dairy is extremely useful in clinical research. It is a more accurate record of actual drug ingestion than memory recall, considering that a patient relying on memory alone to recall drug use will tend to underestimate it, particularly if that drug is an opioid.³⁷

Analgesic consumption has been included in the pain diary in several studies as a method of measuring pain. This method assumes that the patients will take analgesics in the event of pain and that the dose of analgesics taken matches the pain experience. It also assumes that the drug is effective for the pain being experienced. These assumptions cannot be always justified. Because of these assumptions, analgesic consumption, used as a sole means of measuring pain, provides an uncertain assessment of pain.³⁸  It should only be used in conjunction with other instruments.

Pain drawings
A major source of information in assessing pain is the patient’s graphical depiction of the pain (Figure 2). These drawings help identify the location of the pain and the type of pain perceived at the various locations on the body. These pain drawings may help in the diagnosis of the causes of the pain. Organic pain is represented by clearly defined areas that are logical results of potential pathology; for example, the demarcated lancinating pain in the legs that results from a herniated disc compressing a nerve root. Non-organic pain, however, is characterised by diffuse, global, poorly defined patterns that are not a logical result of pathology. Emotional content can also be identified using the pain drawing test. A recent study has demonstrated that patients with clearly abnormal pain drawings are almost always distressed.³⁹ Pain drawings have also been incorporated into many of the multidimensional pain scales.

MULTIDIMENSIONAL PAIN SCALES

The questionnaire is widely applicable. It has been used to measure pain in patients with cancer, chronic pain, to discriminate between different aetiologies of facial pain, to measure dental pain, and to compare the pain experience in acute and chronic pain.⁴¹ The reliability and validity of MPQ has been confirmed in many studies.^42,43 It is used more often as a diagnostic tool rather than as a method for measuring pain intensity in analgesic trials.

Despite its widespread application, the MPQ is time-consuming to perform and does not readily lends itself to the assessment of analgesic effects, where measurements must be polled at 15 to 30 minutes intervals. Perhaps its most serious drawback is its requirement that the patient possess a fairly sophisticated vocabulary. If the patient does not understand the words used, the value of this test is diminished.

Specialised psychological assessment
The role of various psychological/ emotional states in the experience of pain is undeniable. It is clear that the expression and perception of pain is influenced by anger, anxiety, and depression, to name a few. A thorough evaluation of patients with chronic pain must include an assessment of the psychological and social factors associated with their subjective experiences and pain behaviours. The purpose of a psychological evaluation is to frame the pain experience in the context of a patient’s life. Specifically, it evaluates the impact of pain on patient’s functioning and the role that patient’s psychological makeup has in the experience of pain. This assessment usually requires at least one interview and the administration of one or more psychological self-report measures.

Interviews
Interviews must be tailored to fit the situation. However, there is some organisation to the structure. A biopsychosocial format is the most appropriate. Its purpose is to determine how the pain has affected the patient’s life and what factors may influence the pain. There are a few structured behavioural interviews developed for use for chronic pain. However, the most well-known is the psychosocial pain inventory (PSPI).⁴⁴ This measure may be used to gather information from patient and family members of 25 psychosocial aspects of chronic pain. Table 3 summarises the 25 PSPI items, each of which is scored on a 4-point scale using a standardised system. A high score on a PSPI item indicates a greater contribution of that item to the patient’s pain problem.

Psychological self-report measures
Self-report measures of psychological status of chronic pain patients are important for clinical and research purposes because they provide standardised, reliable, and valid assessments of psychological variables that may influence pain perception and behaviour. The self-report measures are usually more sensitive to treatmentrelated changes than are interviews. These may be used by clinicians to guide or evaluate treatment interventions.

Investigation of psychological factors in chronic pain has frequently been explored with the Minnesota multiphasic personality inventory (MMPI); a psychological tool designed to classify patients according to personality types. The test is long and takes highly specialised training to interpret. It may potentially be useful to the physicians if the patient is referred to a psychologist for administration of the test and interpretation of results. Briefly, MMPI consists of 556 items, takes approximately two hours to complete, and provides three validity and ten clinical scales. The clinical scales include: hypochondriasis, depression, hysteria, psychopathic deviance, masculinity-femininity, paranoid, psychasthenia, schizophrenia, hypomania, and social introversion. The three validity scales, listed as L, F, and K, help to indicate whether a patient tries to present fewer or more problems than may be true. Neurotic triad is a combination of scales 1 - 3 of the MMPI. Chronic pain is associated with very high scores on the three scales of the neurotic triad, although scores on the other scales may be within the normal range. Interestingly, the neurotic triad of chronic pain patients showed a typical pattern with a “conversion-V” shaped profile (1,2,3 elevated with two lower than 1 and 3).⁴⁵ Generally, whether using the MMPI or the more recently revised and restandardised MMPI-2, elevations on scales 1,2,3-hypochondriasis, depression, and hysteria- were associated with perceptions of severe pain, affective disturbance, and maladaptive patterns of psychosocial functioning. When elevations on psychopathic deviate and schizophrenia scales were found accompanying elevations in scales 1, 2, and 3, higher levels of psychopathology and resistance to modification of pain behaviour were observed.⁴⁶ 

Symptom checklist 90, revised (SCL-90R) is a self-report measure to quantify the degree of psychological distress.⁴⁷ It assesses nine clinical dimensions:obsessive-compulsive disorder, depression, anxiety, paranoid ideation, somatization, interpersonal sensitivity, hostility, psychoticism, and phobic anxiety. Patients are asked to rate how much they are distressed by each of the 90 described situations. Ratings are quantified on a six-point scale, ranging from zero (not at all) to 5 (extremely). The reliability and validity of the SCL-90R have been demonstrated in a large number of studies.⁴⁸ This test is much briefer than the MMPI and produces much less patient resistance to psychological assessment.

A variety of self-report instruments have been used in the assessment of depression, including the Beck depression inventory (BDI), Hamilton rating scale for depression, and the Zung self rating depression scale.^49,50,51 The BDI is perhaps the most commonly used instrument in the literature. The BDI is a questionnaire consisting of 21 sets of statements. Each set of statements is ranked in terms of severity and scored from 0 to 3. Each set contains statements that express feelings commonly seen in depression. These are issues of guilt, selfworth, and suicidal ideation. It has 10 positive and 10 negative statements. However, patients sometimes find the statements in BDI confusing.

Figure 2a. An organic pain drawing. Note the well localised, anatomically plausible distribution of pain. This is a body diagram prepared by a young lady complaining of bilateral neck, shoulder and arm pain with distal limb numbness. She had bilateral myofascial trigger points in the neck and shoulders with referred pain and paraesthesias in the wrists and hands. Eventual complete resolution of her problem was accompanied by dental treatment of myofascial trigger points of the neck and jaw, and correcting stresses in neck and shoulder muscles created by unusually awkward working postures.

Figure 2b: A non-organic pain drawing. Note the widespread, poorly defined pattern of pain distribution

The functional disability of patients with chronic pain may be assessed with the sickness impact profile (SIP) or the pain disability index.^52,53 SIP is a 136-item measure of pain related functional disability.⁵² The SIP provides a profile of patient disability in 12 dimensions of functioning. These are ambulation, mobility, body care and movement, social interaction, communication, alertness, emotional behaviour, sleep, rest, eat, work, home management, and recreation.

The patient is asked to respond true or false to statements that reflect their impairments in various areas. This instrument is so sensitive that it would detect changes within groups over time, regardless of the medical, demographic, or cultural variables involved. This test can be readministered, yielding a pattern of improvement or deterioration. The major drawback is its length. Patients who are highly disabled often have difficulty in responding to 136 items without assistance.

The patient’s ability to cope with chronic pain can be assessed by the multidimensional pain inventory (MPI) previously known as the West Haven-Yale multidimensional pain inventory or the coping strategies questionnaire.^54,55

The MPI consists of questions regarding family support and interactions, activity level, and pain intensity. Nine clinical scales encompass ratings of pain severity, interference of pain, life control, affective distress, social support, punishing responses, and general activity level. Probabilities are developed of the patient matching one of the three profile types: dysfunctional-characterised by high pain level, decreased activity level and sense of control, and increased emotional distressed, interpersonally distressed patients perceiving themselves as lacking in support from family and friends, and adaptive coper patients characterised by reporting low level of pain, interference, and interpersonal support.

Physiological measures
The advantages of physiological correlates are manifold. To the clinician, it represents an “objective” measure of this subjective experience. To the scientist, it represents the search for the biological bases of pain perception and its physiological expression. However, it is still limited because pain is so inextricably bound up with the emotional context within which it is perceived that its biological separation at this level is probably not meaningful.

Physiological manifestation
Heart rate initially increases with acute, sharp pain but later decreases. Arterial oxygen tension decreases during painful procedures and stress-related hormones are released. However, fear alone produces similar changes in oxygen saturation, and blood hormone levels depend on age, diurnal rhythms, emotions, and baseline values. Evoked potentials recorded from the scalp have been shown to be linked with the intensity of a stimulus. However, no single parameter clearly distinguishes a painful from a non-painful stimulus. In contrast to acute pain, there are few useful physiologic indicators of chronic pain. Although, these physiological measures are more objective, in many cases they are not reliable indices of pain.

Biochemical correlates
Various biochemical mediators have been shown to be related to pain. However, the more relevant data come from the acute pain model. Recently, a breakthrough study has been conducted. A microdialysis technique has been developed that permits collection of inflammatory mediators in conscious dental surgical patients, who can report pain scores simultaneously.^56-58 These clinical studies have used the microdialysis probes to collect tissue levels of immuno-reactive bradykinin, prostaglandin E₂, leukotriene B4, substance P, and other inflammatory mediators. It indicates that all the above mediators are detectable in tissue dialysates collected from extraction site of awake patients after surgical removal of third molars. Data for time of peak levels and peak concentrations of the mediators were determined and have been shown to correlate well with peak pain scores. Pharmacological studies have also evaluated the effects of non-steroid antiinflammatory drugs (NSAIDs) and steroid treatment on tissue levels of inflammatory mediators in this model. As compared with placebo treatment, administration of NSAIDs and steriod significantly reduces tissue levels of these mediators and this correlates with a reduction in pain.^56-58

ANALYTIC APPROACHES FOR INTERPRETING SUCH DATA
Measure of pain and its relief, obtained by VAS, NRS or VRS scales, are analysed by parametric analyses in some published studies and by nonparametric procedures in others. The selection of method of analysis is based on the assumptions that are made about the data collected. For example, should categorical ratings of pain relief be considered as representing an interval scale, and meeting the properties of that level of measurement, or should these evaluations be considered as satisfying the less stringent criteria of an ordinate measure? In general, parametric analyses have more power, that is, the ability to detect a real difference between treatments, than nonparametric procedures if the data meet the necessary assumptions. On the other hand, nonparametric procedures generally give true alpha levels under very broad assumptions about the underlying variables. It has been recommended that the data should be analysed using both parametric and nonparametric procedures.⁵⁹

Analgesic efficacy variables are usually derived from pain intensity and pain relief data. Pain relief can be calculated as pain intensity difference (PID), i.e., the difference at a certain time of assessment from a starting pain, and over an observation period as the sum of pain intensity difference (SPID) scores. Patients can also be instructed to estimate the degree of pain relief (PAR) from the initial pain. Over an observation period the total pain relief (TOTPAR) can be calculated as the sum of PARs. However, some suggest that a more appropriate measure of change is obtained by having patients rate to the absolute amount of pain at different points of time before and after an intervention, since patient assessment of pain relief (PAR) may introduce unnecessary bias.⁶⁰ For both measures of pain intensity and relief, serial visual analogue scores can be complied into a graph of either pain intensity or pain relief. The area under the curve (AUC) can be calculated using the trapezoidal method to provide an integrated measure of pain intensity (AUC intensity) and pain relief (AUC relief) throughout the observation period.⁶¹

FACTORS AFFECTING PAIN MEASUREMENT
Research from a number of sources and with different populations provides strong evidence that many factors influence communication about pain. Factors like the environment, patient’s/clinician’s belief, placebo effect,cultural background, age, and gender roles may affect pain report. The clinician often expects a high correlation between the patient’s self-report of behaviour and observed behaviour. There is often, however, marked disparity when comparing a patient’s own assessment of acute pain with physician ratings of the same patient’s pain experience. A brief review of this research will help to illustrate this important point.

In one study, college students were assigned to be in the presence of an attractive male or female while holding their hand in ice water and rating pain.⁶² Based on gender role expectations, the authors predicted that males would report less pain in the presence of the female as opposed to the male experimenter. This is exactly what they found. Female students, on the other hand, were not significantly influenced by the gender of the experimenter. In another study, it was found that modelling of pain tolerance impacts the report of pain.⁶³ Subjects who observed people modelling high pain tolerance reported higher pain tolerance than subjects who observed people modelling low pain tolerance. Finally, it was shown that the changing environment influenced pain report. The subjects reported that tooth pulp stimulation hurt more when they were administered in a dental clinic than when they were administered in a research laboratory setting.⁶⁴

Placebo effect may affect pain report and there are problems of identifying potential placebo responses. The patient’s mental state at the time of the pain experience will affect his or her interpretation of and reaction to that pain experience. The clinician can strongly influence a patient’s pain intensity rating by antecedent reinforcement of pain talk or well talk.⁶⁵ The length of a clinical trial is also important since it can influence apparent analgesia. Total analgesic effect increases for placebo as the number of hours of evaluation increases; this appears to be true for all measures of total effect. This exaggerated placebo effect may obscure differences between placebo and weak analgesic standards, such as aspirin 650mg or codeine 60mg, in 8 or 12 hour studies.⁶⁶

There is a possible influence of cultural background in pain reports.⁶⁷ Higher pain threshold to calibrated noxious stimuli has been reported more among Northern Europeans than among Mediterranean people and African Americans. Irish Catholics and Yankee Protestants have been reported to have higher pain thresholds than Italians and Jews. However, as summarised in a recent review, there is little evidence for ethnocultural differences in the discrimination of noxious stimuli, but there are significant cultural differences in reporting pain.⁶⁷ Probably all the differences in pain thresholds that have been reported among various ethnocultural and religious groups are due entirely to cultural differences in the criterion for reporting pain and not at all to differences in the sensory experience of pain itself.

Apparent gender differences have been identified in epidemiological surveys of patients with pain and in clinical studies of responses to pain. Women are reported to have lower pain thresholds and lower pain tolerance than men have. Whether women are more willing to report pain than men are or experience pain differently than men do is unclear.⁶⁸ However, beliefs about gender differences and pain affect decisions made regarding the treatment of pain. Pain experiences may also vary greatly with patient’s age.⁶⁹ Some studies indicate that pain threshold and pain tolerance with heat and pin prick stimuli increase with age. This may be to possible changes in receptors or nervous system structures and because enhanced maturity affects attitudes about pain.

One of the implications of this body of research is that clinicians and researchers should take the factors known to influence pain measurement into account when assessing pain. The conditions under which the report of pain is made should be as similar as possible between comparison groups or between assessment periods. Attempts should be made, for example, to have patients rate their pain at the same time of the day, in the same place, and in the presence of the same people at each assessment period. Aggregating multiple measures, e.g., taking the average of several pain measures may help to minimise the influence of extraneous or irrelevant contextual factors. In support of this practice, aggregated pain measures have been shown to be more sensitive to treatment effects than single measures.

RECOMMENDATIONS TO GOOD CLINICAL PAIN MEASUREMENT
In selecting a pain measure for clinical or research use, one must ensure that the measure is reliable and valid for the patient’s group chosen and practical in the clinical situation and that the measure is appropriate for the type of pain being assessed (e.g., acute versus chronic). No single pain measurement instrument can provide all the required information. In general, the plain VAS is probably the most reliable and sensitive tool for measuring pain. MPQ is also a very reliable and valid tool especially for measuring chronic pain. These two instruments appear to fulfil most of the properties of an ideal pain measurement system proposed by Gracely and Dubner (1981).¹¹ In addition, pain assessment can be enhanced by comparing self-report measures with behavioural and physiological measures in certain specific situations.

It is important to have an organised approach to the measurement of pain. An algorithm for the selection of pain measurement instruments is shown in Figure 3. Almost all patients can use at least one of the single dimension pain measurement instruments. These instruments are easily understood by the patient, easily administered by the clinician, display appropriate reliability and validity, and produce results that can be used in assessing analgesic efficacy. Many patients can also understand the multidimensional tests, but their administration can be more time consuming than single dimension tests. Behavioural testing is appropriate in many patients, especially those with a possible functional component of their pain.

Figure 3: An algorithm for selection of pain measurement instruments. BDI- Beck Depression Inventory, CSQ- Coping Strategies Questionnaire, MPQ- McGill Pain Questionnaire, MPI- Multidimensional Pain Inventory, MMPI- Minnesota Multiphasic Personality inventory, NRS- Numerical Rating Scale, SCL-90R-Symptom Checklist List 90, Revised, SIP- Sickness Impact Profile, PDI- Pain Disability Index, VAS- Visual Analogue Scale, VRS- Verbal Rating Scale.

A strong recommendation is to keep pain measurement scales simple when assessing subjective pain reports. For quality control, more than one measurement tool should be used and two or three dimensions in any given situation is recommended, especially in chronic pain. In general for acute and brief pain, use rapidly administered self-report methods and observations, such as VAS, NRS or VRS and behavioural observations. For chronic and persistent pain, assess the patterns of pain over time along several dimensions, such as VAS and MPQ, and other psychological and behavioural observation. Particularly for long-term complex chronic pain, a battery of assessment tools may be required. A detailed interview and some measure of personality style can be done to gauge permorbid adaptation as well as current tendencies with tools like the MMPI. These will provide some evidence of the affective disturbance in addition to the pain severity. Specific measures for associated emotions like depression, anxiety, and anger can be useful in certain cases. Scales such as BDI, STAS and SCL-90R are reliable and can be used. It is important to get a sense of the patient’s perceived level of coping, including some measure of perceived disability; the MPI and SIP could supply this information. No assessment would be complete without a daily pain diary, which records a patient’s daily activity level, medication usage, pain intensity, and mood over time. Such a comprehensive battery can be very helpful in determining the course of treatment and as a baseline against which to measure outcome. Due to the complex nature of these tests, the help of a psychologist and other allied health professionals, is usually required for the comprehensive assessment.

However, it must be remembered that pain measurement per se cannot be used as an endpoint. Clinical correlation and individualised treatment must be used to guide the clinician for good pain management.

REFERENCES
1. Merskey H, Bogduk N, editors. Classification of chronic pain - Descriptions of chronic pain syndromes and definitions of pain terms, 2nd ed. IASP press, Seattle, 1994: 40-43.
2. Melzack R, Casey KL. Sensory, motivational, and central control determinants of pain. A new conceptual model. In: Kenshalo DR, (ed). The Skin Senses. Thomas, Springfield, 1968, 423-443.
3. Loeser JD. Perspective on pain. In Turner P (ed): Proceedings of First World Congress on Clinical Pharmacology and Therapeutics. Macmillan, London, 1980: pg 316.
4. Gracely RH. Evaluation of multidimensional pain scales. Pain 1992; 48: 297-300.
5. Grossman SA. Assessment of cancer pain: a continuous challenge, support. Care Cancer 1994; 2: 105-110.
6. Dubuisson D, Melzack R. Classification of clinical pain descriptions by multiple group discriminant analysis. Experimental neurology 1976; 51: 480-487.
7. Max MB, Portenoy RK, Laska EM (eds). The Design of Clinical Analgesic Trials, Raven Press, New York, 1990.
8. Chapman CR, Loeser JD (eds). Issues in Pain Measurement, Raven Press, New York, 1990.
9. Osterweis M, Kleinman A, Mechanic D. Pain and disability: Clinical, Behavioural and Policy Perspective, National Academy Press, Washington DC, 1987.
10. Vasudevan SV. Impairment, disability, and functional capacity assessment. In Turk DC, Melzack R, (eds): Handbook of Pain Assessment. Guilford Press, New York, 1992: pg 100.
11. Gracely RH, Dubner R. Pain assessment in humans- A reply to Hall. Pain 1981; 11: 109-120.
12. Turk DC, Melzack R. The measurement of pain and the assessment of people experiencing pain. In Turk DC, Melzack R, (eds): Handbook of Pain Assessment. Guilford Press, New York, 1992: Pg 3-14.
13. Jensen MP, Karoly P. Self-report Scales and procedures for assessing pain in adults. In Turk DC, Melzack R, (eds): Handbook of Pain Assessment. Guilford Press, New York, 1992: Pg 135-151.
14. Seymour RA. The use of pain scales in assessing the efficacy of analgesics in postoperative dental pain. Eur J Clin Pharmacol 1982; 23: 441-444.
15. Gracely RH, McGrath P, Dubner R. Validity and sensitivity of ratio scales of sensory and affective verbal pain descriptors: Manipulation of affect by diazepam. Pain 1978b; 5: 19-29.
16. Jensen MP, Karoly P, breaver S. The measurement of clinical pain intensity: A comparison of six methods. Pain 1986; 27: 117-126. 
17. Jensen MP, Karoly P, O’Riordan EF. The subjectivity experience of acute pain: An assessment of the utility of 10 indices. Clin J Pain 1989; 5: 153-159.
18. Chapman CR, Donaldson GW, Jacobson RC. Measurement of acute pain states. In Turk DC, Melzack R, (eds): Handbook of Pain Assessment. Guilford Press, New York, 1992: Pg 332-343.
19. Wallenstein SL, Heidrich G III, Kaiko R. Clinical evaluation of mild analgesics: The measurement of clinical pain. Br J Clin Pharmacol 1980; 19: 319S-327S.
20. Keefe FJ, Schapira B, Williams RB. EMG-assisted relaxation training in the management of chronic low back pain. Am J Clin Biofeedback. 1981; 4: 93-103.
21. Farrar JT, Young JP, LaMoreaux L. Clinical importance of changes in chronic pain intensity measured on an 11-point numerical pain rating scale. Pain 2001; 94: 149-158.
22. Turner JA. Comparison of group progressiverelaxation training and cognitive-behavioural group therapy for chronic low back pain. J Consult Clin Psychol 1982; 50: 757-765.
23. Chambers CT, McGrath PJ. Pain measurement in children. In Asburn MA, Rice LJ (eds): The management of PAIN, Churchill Livingstone, New York, 1998, 625-634.
24. Joyce CRB, Zutshi DW, Hrubes V. Comparison of fixed interval and visual analogue scales for rating chronic pain. Eur J Clinical Pharmacol 1975; 8: 415- 420.
25. Downie WW, Leatham PA, Rhind VM. Studies with pain rating scales. Ann Rheumatic Disease 1978; 37: 295-303.
26. Price DD, Bush FM, Long S, Harkins SW. A comparison of pain measurement characteristics of mechanical visual analogue and simple numerical rating scale. Pain 1994; 56: 217-226.
27. Max MB, Shafer SC, Cilane M, Dubner R. Association of pain relief with drug side effects in post-herpetic neuralgia: A single dose study of clonidine, codeine, ibuprofen, and placebo. Clinical Pharmacol Therapeutics 1987; 43: 363-371.
28. Scott J, Huskisson EC. Graphic representation of pain. Pain 1976; 2: 175-184.
29. Carlsson AM. Assessment of chronic pain: aspect of reliability and validity of the visual analogue scale. Pain 1983; 16: 87-101.
30. Walsh TD. Practical problems in pain measurement. Pain 1984; 19: 96-98.
31. Gracely RH, McGrath F, Dubner. Ratio scales of sensory and affective verbal pain descriptors. Pain 1978; 5: 5-18.
32. Gracely RH, Kwilosz DM. The Descriptor Differential Scale: applying psychophysical principles to clinical pain assessment. Pain 1988; 35: 279-88.
33. McGrath PA, Brigham MC. The assessment of pain in children. In Turk DC, Melzack R (eds): Handbook of Pain Assessment, The Guilford Press, New York, 1992, 295-314.
34. Craig KD, McMahon RJ, Morison JD, Zaskow C. Developmental changes in infant pain expression during immunization injections. Social Science & Medicine 1984; 19: 1331-1337.
35. Frank AJM, Moll JMH, Hort JF. A comparison of three ways of measuring pain. Rheumatology & Rehabilitation 1982; 21: 211-217.
36. Ekman P, Friesen W. A new pan-cultural facial expression of emotion. Motivation & Emotion 1986; 10: 159-168.
37. Salovey P, Smith AF, Turk DC, Jobe JB. The accuracy of memory for pain: Not do bad most of the time. APS J 1993; 2: 184-191.
38. Seymour RA. The Measurement of Pain. Dental Update 1983; 446-454.
39. Parker H, Wood PLR, Main CJ. The use of pain drawings as a screening measure to predict psychological distress in chronic low back pain. Spine 1995; 20: 236-243.
40. Melzack R. The McGill Pain Questionnaire: Major properties and scoring methods. Pain 1975; 1: 277-299.
41. Chapman CR, Casey KL, Dubner R, Foley KM, Gracely RH. Pain management: An overview. Pain 1985; 22; 1-31.
42. Reading AE. An analysis of the language of pain in chronic and acute patient groups. Pain 1982; 13: 185-192.
43. Love A, Loeboeuf DC, Crisp TC. Chiropractic chronic low back pain sufferers and self-report assessment methods. A reliability study of the visual analogue scale, the pain drawing and the McGill Pain Questionnaire. J Manipulative Physiological Therapeutics 1989; 12: 21-25.
44. Geto CJ, Heaton RK, Lehman RA. PSPI: A standardised approach of to the evaluation of psychosocial factors in chronic pain. Advances in Pain Research & Therapy 1983;5: 885-889.
45. Michelotti A, Martina R, Russo M, Romeo R. Personality characteristics of temporomandibular disorder patients using MMPI. CRANIO 1998; 16: 119-125.
46. Bradley LA, McDonlad HJ, Jaworski TM. Assessment of psychological status using interviews and self-report instruments. In: Handbook of pain assessment. Turk DC, Melzack R (eds). New York, Guilford Press, 192-213.
47. Bernstein IH, Jaremko ME, Hinkley BS. On the utility of the SCL-90R with low back pain patients. Spine 1994; 19: 42-48.
48. Spinhoven P, Linssen ACG. Behavioural treatment of chronic low back pain: I. Relation of coping strategy use to outcome. Pain 1991; 45: 29-34.
49. Beck AT, Ward CH, Mendelson M. An inventory for measuring depression. Archives General Psychiatry 1961; 4: 561-571.
50. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960; 23: 56-68.
51. Zung WWK. A self-rating depression scale. Archives General Psychiatry 1965; 12: 63-70.
52. Follick MJ, Smith TW, Ahern DK. The Sickness Impact Profile: A global assessment of disability in chronic low back pain. Pain 1985; 21: 67-76.
53. Tatt RC, Chibnall JT, Krause S. The Pain Disability Index: Psychometric properties. Pain 1980; 40: 171-182.
54. Kerns RD, Turk DC, Rudy TE. The West Haven-Yale Multidimensional Pain Inventory. Pain 1985; 23: 345-356.
55. Rosential AK, Keefe FJ. The use of coping strategies in chronic low back pain patients: Relationship to patient characteristic to current adjustments. Pain 1983; 17: 33-34.
56. Hargreaves KM, Costello A. Glucocorticoids suppress release of immunoreactive bradykinin from inflamed tissue as evaluated by microdialysis probes. Clin Pharmacol Ther 1990; 48: 168-178.
57. Swift JQ, Garry MG, Roszkowski M, Hargreaves KM. Effect of flubiprofen on tissue levels of immnureactive bradykinin and acute post-operative pain. J Oral Maxillofac Surg 1993; 51: 112-116.
58. Roszkowski M , Swift JQ, Hargreaves KM. Prostaglandin E2 tissue levels increase following third molar extraction. J Dent Res 1992; 71: 178.
59. Forbes JA. Oral Surgery. In Max MB, Portenoy RK, Laska EM (eds). The Design of Clinical Analgesic Trials, Raven Press, New York, 1990, pg 347-374.
60. Melzack R, Katz J. Pain measurements in persons in pain. In Wall PD, Melzack R (eds). Textbook of Pain. 3rd ed. Edinburgh; Churchill Livingstone, 1994, pg 337-351.
61. Matthews JNS, Altman DG, Campbell MJ, Royston P Analysis of serial measurements in medical research. BMJ 1990; 300: 230-235.
62. Levine FM, De Simone LL. The effects of experimenter gender on pain report in male and female subjects. Pain 1991; 44: 69-72.
63. Craig KD, Weiss SM. Vicarious influences on pain-threshold determinations. J Personality Social Psychology 1971; 19: 53-59.
64. Dworkin SF, Chen CAN. Pain in clinical and laboratory contexts. J Dent Res 1982; 6: 772-774.
65. Gracely RH, Dubner R, Deeter WR, Wolskee PJ. Clinicians' expectations influence placebo analgesia. Lancet 1985 Jan 5; 1(8419): 43.
66. Stewart PE. Multidose short-term analgesic studies. In Max MB, Portenoy RK, Laska EM (eds). The Design of Clinical Analgesic Trials, Raven Press, New York, 1990, pg 137-150.
67. Zatzick DF, Dimsdale JE. Cultural variations in response to painful stimuli. Psychosom Med 1990; 52: 544-557.
68. Cook AJ, Chastain DC. The classification of patients with chronic pain: Age and sex differences. Pain Res Manag 2001; 6: 142-151.
69. Woodrow KM, Friedman GD. Siegelaus AB. Pain defferences according to age, sex, and race. Psychosom Med 1972; 39: 548-556.

Copyright: 9 January 2004