Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
11 Measuring Pain and Dysfunction This chapter explores the technical feasibility of measuring chronic pain and related dysfunction. It addresses questions of whether the severity of pain can be docu- mented and whether the relation between pain and the ability to work can be assessed. The Social Security Administration (SSA) considers these questions to be crucial to disability evaluation of pain patients. Physiological and neurological techniques for measuring pain were described in Chapter 7; this chapter examines psychological, behav- ioral, and functional methods for assessing pain. Measures of con~- tions such as depression and anxiety are not discussed because these have been thoroughly reviewed elsewhere (see Anastasi, 1983; Evans, 1983; L`ehmann, 1985; Spielberger et al., 19841. Some of the methods discussed here are appropriate for the assessment of work disability for compensation purposes; others are more appropriate for use by clini- cians diagnosing and treating pain patients. In order to pursue its mandate to conduct a state-of-the-art evalua- tion of pain assessment methods, the Institute of Medicine (IOM) study committee reviewed the literature and invited a panel of six experts in pain measurement and related matters to make formal presentations and participate in a discussion at one of its meetings (see list of panelists at end of chapter). Topics reviewed with the panelists included physiological measures of pain, subjective and behavioral observation techniques, assessment of the meaning of pain, physical function measures and vocational assessment in chronic pain patients, and the assessment of psychosocial and psychiatric factors in the 211
212 ASSESSING ED TREATING PMN ED DYSFUNCTION etiology and maintenance of chronic pain and dysfunction. In later meetings the issues of assessing pain and functional capability were discussed repeatedly and considered in light of disability evaluation. BASIC CONCEPTS OF MEASUREMENT Measurement is the process of assigning numbers to specific proper- ties of events, processes, objects, or persons. All measurement involves error to varying degrees. Number assignment, or scaring, may involve different degrees of precision. It can tee used to categorize or classify items or individuals (e.g., 1 = male, 2 = femaTe) or to rank order them. In some cases, number assignment is sufficiently precise to justify mathematical manipulation of scores. The depth of scientific inference permitted by measurement depends on the precision of the scaling used. In some cases the attributes or properties being measured are hypothetical, being derived from a theory or model, rather than an object or event that can be objectively observed. Human pain, like intelligence, cannot be directly observed but may be scaled along one or more dimensions in accordance with theory or highly specific models. Pain measurement is never Theoretical; every too! is rooted in a fundamental conceptualization of pain or at least certain basic assumptions that may or may not be explicitly defined by its users. There are both medical and nonmedical conceptual models for pain. In each of these broad categories, there are several subcategories. Medical moclels share the basic assumption that pain is a symptom of an underlying pathology. Nonmedical models, which are largely psy- chological in nature, construe pain as a perception, as a behavior, or as . . a cognition. The distinctions between the various models are complicated by the basic distinction between acute and chronic pain. Some models can account for one type of pain but not the other. In general, medical models are best suited for explaining acute pain problems in which pain is a direct fimction of nociception, whereas nonmedical models can best account for chronic pain problems in which the relation between tissue damage and pain complaint is weak or lacking. In practice, medical and nonmedical models can be, and often are, combined. The multiplicity of models available for quantifying pain attests to the early stage of development of this area of research. The lack of a unified theoretical perspective has both advantages and disadvantages and is probably a necessary stage in the long-range development of science in this difficult area. On the positive side, the combination of
MEASURING PAIN AND DYSFUNCTION 213 medical and nonmedical models is a more powerful approach to pain assessment and control than the use of either model alone. On the practical, clinical level, measures derived from multiple models may provide a mosaic of information that promotes a clearer understanding of the patient than the medical model alone. The primary disadvantage of variation in theoretical perspective is inconsistency in the resulting measurement technology. Operational definitions of pain vary greatly, and the data gathered by one inves- tigator may be of little or no use to another. In addition, investigators sometimes disagree with one another about precisely what is being measured. For example, as discussed in Chapter 9, pain may be a symptom of depressive disorder. a consequence of it, or a problem that coexists with depression. Depression, then, may be an independent variable for one theorist and a dependent variable for another. The lack of consensus on the role of affect in chronic pain is a major impediment in the progress of pain measurement technology. MEASUREMENT TOOLS Most pain measurement involves either structured self-report of pain, observation of patient behavior, or some combination of the two approaches. When self-report methods are used, instruments should (1) burden patients minimally, (2) be understood by patients, (3) yield a wide range of scores with sensitivity to analgesic intervention, and (4) demonstrate appropriate reliability and validity. Observational methods for the scaling of pain need not be understandable to the patient, but they must fulfill the other three criteria and, in addition, protect the patient's right to privacy. The major methods for scaling pain and their advantages and limitations are discussed below. Measurement of Subjective Pain States Measurement by subjective report is by far the most common type of procedure for quantifying pain. Patients may indicate pain levels verbally, mark simple scales, or fill out complex questionnaires. In all cases, the patient determines the data. Self-Report: Data from Introspection Many investigators hold that pain is inherently a private experience that can only be quantified by asking the patient to do his or her own number assignment. There are both unidimensional, or simple, scales
214 ASSESSING ED TREATING Ply ED DYSFUNCTION tor pain measurement and multidimensional, or complex, methods to obtain data. Self-Report Methods Using a Single Dimension The simplest approach to assessing subjective pain states is the use of category scales. Usually the intensity dimension is scaled, but category methods can scale aversiveness or some other quality of pain as well. Such scales require only simple choices of the best descriptors from the patient. For example, MeIzack and Torgerson (1971) intro- duced the following scale for pain intensity: "mild, discomforting, distressing, horrible, excruciating." Statistical treatment of category data is usually restricted to nonparametric methods, and this restrains the interpretation of the data gathered. Moreover, respondents tend to use the middle of the scale. An alternative is the Visual Analog Scale (VAS), which usually consists of a 10-cm line anchored at one end by a label such as "no pain" and at the other end by "the worst pain imaginable" or "maximum pain." Respondents mark the line to indicate pain intensity; the mark is scored on either a 1-10 or a 1-100 scale. The Numerical Rating Scale (NRS) is a variation of the VAS in which patients rate their pain on a ~10 or ~100 scale that is discrete rather than continuous. Although these tools are expedient, researchers have repot ted that 7 to 11 percent of patients are unable to complete the VAS or find it confusing and about 2 percent are not able to use the NRS (Kremer et al., 1980; Revill et al., 1976; Walsh, 19841. Carisson (1983) critically reviewed the VAS as a method for scaling pain states or pain relief in chronic pain patients, and she compared different forms of the scale. Reliability, as judged from consistency of responses to two forms, was Tow, and CarIsson concluded that the validity of VAS procedures for chronic pain populations may be unsatisfactory. The VAS is a straightforward, efficient tool for scaling pain, but it can fad] if care is not taken to ensure accurate, valid, and reliable reporting. Such instruments will continue to be used because of their expediency, minimal respondent burden, and face validity. However, in addition to the limitations in reliability, these tools may oversim- plify the pain experience. Self-Report of Pain in Several Dimensions The scientific value of the VAS is restricted by its unidimensional- ~ty. Some investigators adjust for this by using more than one VAS,
MEASURING PEN ED DYSFUNCTION 215 with each designed to assess a different dimension of pain. However, it is hard to ensure that the response to the first scale a~ninistered does not influence the response to subsequent scales (Carisson, 19831. There are several multidimensional scales that avoid some of the problems with VAS but increase responder burden and cost more to interpret. The McGill Pain Questionnaire (MPQ) is perhaps the most thor- oughly evaluated multidimensional scaling device for pain. It is based on the vocabulary used by patients to describe various experiences of pain. The MPQ scales pairs along three dimensions: sensory, affective, and evaluative. There are 20 sets of words that describe varying qualities of pain. Ten of the sets represent sensory qualities, 5 are affective, and 1 is evaluative. Each set has from two to six words that vary in intensity for the quality described by the set (e.g., from hot to searing, from annoying to unbearable). Patients are instructed to select the sets that are relevant to their pain and to circle the words that best describe it within each selected set. An adjunct test device, the Dartmouth Pain Questionnaire, has been offered to supplement the MPQ through the assessment of four additional factors, including impaired functioning (Corson and Schneider, 1984~. Many studies support the factor structure of the MPQ, its reliability, and its concurrent validity (see Shale and Chapman, 1984; Chapman et al., 1985~. However, it places a large responder burden on the patient, some patients cannot handle the vocabulary of the instru- ment, and the scoring procedures available are limited (Sy~ala and Chapman, 1984~. Turk, Rudy, and Salovey (1985) critically evaluated the MPQ and various approaches to scoring it. They concluded that the total score is valid as a general measure of pain severity but that individual scale scores should not be used; adequate discriminant validity to support scaling at the level of sensory, affective, and evaluative dimension subscales could not be demonstrated. Meizack has responded to these challenges by pointing out that the high intercorrelation among the factors of the MPQ does not necessarily indicate a lack of discriminant validity and by reviewing an impressive number of studies that demonstrate the discriminant capacity of the instrument (Meizack, 1985). The West Haven-Yale Multidimensional Pain Inventory is an alter- native instrument designed to be briefer and more classical in its psychometric approach to multidimensional scaling than the MPQ (Kerns et al., 19853. The 52-item inventory is divided into three parts: (1) five general dimensions of the experience of pain and suffering, interference with normal family and work functioning, and social
216 ASSESSING~D TREATING PANNED DYSFUNCTION support; (2) patients' perceptions of the responses of others to displays of pain and suffering; and (3) the frequency of engagement in common daily activities. The instrument is derived from cognitive-behavioral theory and assesses such constructs. As such, it represents a very different approach to scaling than the MPQ. The Brief Pain Inventory is an efficiently administered multidimen- sional pain questionnaire with demonstrated reliability and validity in cancer and arthritis patients (Daut et al., 1983; Cleeland, 19855. In less than 15 minutes, patients can report analgesic medication use, pain relief from drugs, beliefs about the cause of pain, qualitative descrip- tions of pain, areas in which pain interferes with quality of life, the pain's locus, and their worst, average, and current pain level on a 0~10 scale. This measure has proved useful as a multidimensional pain measure in patients with progressive disease. Other approaches to the multidimensional scaling of pain have been derived from a psychophysical technique known as cross-modality matching in which a sensory experience is quantified by matching it to the experience of a precisely controlled stimulus in a different sensory modality. For example, a laboratory subject might match the intensity of a toothache produced by electrical tooth shock to the loudness of a controlled tone in decibels. Analogous methods are used for clinical pain scaring. The typical procedure involves matching words describ- ing pain to line length or handgrip force, matching both to experimen- tal pain, and then deriving scaling standards for the relationship of words describing pain to actual pain. The methods can then be applied to clinical pain assessment. This can be done for multiple dimensions of pain, such as intensity and unpleasantness (Gracely et al., 19791. The Pain Perception Profile (Tursky et al., 1982) uses cross-modaTity matching scaling procedures. It (1) quantifies the sensation threshold; (2) uses magnitude estimation procedures to judge induced pain; (3) scales pain on intensity, reaction, and sensation dimensions using psychophysical scaling of verbal pain descriptors; and (4) permits the psychophysically scaled verbal descriptors to be used in a diary format for repeated assessment over time. Compared to the MPQ, cross-modality scaling methods are shorter and less demanding. Yet they over potentially more reliable and valid data than the simpler VAS scales. However, the Pain Perception Profile and Graceley's methods (Gracely et al., 1979) have not yet been validated for different patient populations. Such work would require experimental pain testing and a substantial amount of development before clinical data from a broad sample of patients could be inter- preted confidently.
MEASURING PEN ED DYSFUNCTION 217 Pain as Behavior: Observational Data Pain Judgments by Health Care Providers Health care providers generally use two types of pain assessment that can be more or less structured: patient classification and observer ratings of patients' pain problems. The most basic form of observer scaling is patient classification. Typically, medical history and diag- nostic data are used to categorize chronic pain patients. For example, Hammonds and Brena (1983) devised a four-category classification scheme for patients in which Class ~ consisted of those with high behavioral determinants and low organic determinants of pain, and Class lI patients were Tow both on behavioral and organic determi- nants. Class ITI patients had high scores both on organic pathology and pain behaviors, and Class IV patients had high organic pathology and Tow pain behavior scores. Advocates of such approaches point to the value of such categorizations for screening and selecting appropriate interventions. Others, however, believe that such methods are over- simplified, that most patients do not fit neatly into a category, and that such categorization may affect the patient's care inappropriately. When pain is directly scaled by health care providers observing the patient, simple rank-ordered category scales are typically used, such as no pain, slight pain, moderate pain, and severe pain. Often, such scales may include an evaluation of what the patient can or cannot do on certain tasks, such as bending over to pick up a weight. There are several limitations of observer pain judgments in chronic pain settings. First, most investigators hold that such judgments are not true measurements: Pain can never be observed by another individual, it can only be inferred from a patient's actions. Second, knowledge of a patient's clinical findings can bias the pain rating. There is also the danger that raters will stereotype patients on the basis of age, sex, or race. Use of multiple raters who are trained with well-defined criteria for assessing pain can reduce these problems. Measurement of Pain Behavior Although pain may not be objective, the behaviors of patients in pain may be observed, and scored, objectively. Pain behavior is a normal response to an injurious stimulus, but when it occurs in the absence of such a stimulus or too small a stimulus, it may be described as an abnormal behavior. Behaviorists generally do not seek to infer pain from behavior; instead, they view pain behavior itself as the problem;
218 ASSESSING ED TREATING PEN ED DYSFUNCTION and it is pain behavior, rather than some hypothetical personal state, that they seek to correct therapeutically. Patients in pain exhibit certain consistent behavior patterns. For example, studies have shown that back pain patients tend to grimace, guard their movements, rub themselves, and sigh (Keefe and Hill, 19851. When clinical pain is operationally defined in terms of such behavior patterns, objective measures can be used. Certain behavior patterns can be identified as being related to pain, quantified in terms of frequency or rate of occurrence, and assessed via direct or videotape observation of patients in selected settings performing specific tasks. Because behavioral patterns are complex, most behavioral measures are multidimensional. The identification and quantification of pain behaviors varies greatly with different types of pain problems. In general, behaviors are tallied over time and scored in terms of their frequency. Keefe and Block (1982) developed an observational scoring system for scoring pain behavior in chronic back pain patients. Guarded movement, bracing, rubbing, and sighing were assessed. These indices proved reliable, valid (in relation to reported pain), and more frequent in pain patients than in normals or depressed controls. Keefe and Hill (1985) extended the observational approach by using a transducer placed in the patients' shoes so that walking parameters could be assessed. Patients and nonpatients were required to walk a 5-m course while being videotaped. Patients walked more slowly than normals, took smaller steps, failed to show normal symmetrical gait patterns, and exhibited more pain behaviors. This approach appears promising for assessing and objectively quantifying back pain behav- iors. More generally, repeated recordings could be used to document changes in an individual patient's behavior over time for use by physicians or by disability examiners. In the latter instance, such recordings could provide some information that would otherwise be available only by a face-to-face encounter. A major limitation of the behavioral approach is that pain behaviors are highly specific for each pain syndrome. Patients with shoulder pain or headache, for example, would probably be indistinguishable from healthy people on Keefe and Hill's (1985) test. Keefe and his colleagues (1985) undertook behavioral evaluations of patients with head and neck pain to address this issue. They found that such patients dis- played their pain primarily through their facial expression rather than through guarded movements. Broad indicators of pain behaviors with potential application to different clinical populations do not appear promising. Linton (1985) hypothesized that reported pain intensity is inversely related to
MEASUP`ING PAIN AND DYSFUNCTION 219 general activity level in back pain patients (as measured by self- monitoring or observed behavior in a test situation). He found no relation between the level of patient activity and chronic pain inten- sity. Thus, behavioral indicators appear well suited as highly specific and precise ways to quantify certain outward pain behaviors, but they may not serve well as global indicators of subjective pain intensity. Further work is needed before the behavioral approach can quantify a wide range of chronic pain problems. Mixed Methods: Self-Reports of Pain Behaviors One way to gather information about behavior patterns and habits is to ask the patient, his or her spouse, or some other day-to-day observer. The pain diary is perhaps the most commonly used form of behavioral self-report. A typical pain diary is a Tog of daily pain-relevant activity broken down into small blocks of time. Activities may be divided into sitting, walking, and reclining, with the patient filling in the specific activity under the appropriate category according to the time the activity occurred. Pain ferret is rated on a ~10 scale for each hour, and medications may also be recorded. There are several potential advantages to using a pain diary with chronic pain patients. First, because the diary is completed daily, it is not subject to distortion based on the patient's current pain experience at the time he or she is seen in the clinic. Variations in pain levels during the day and from day to day can be recorded. Second, the diary yields data on patterns of normal activity relative to patterns of pain behavior (or pain-linked inactivity) not available from other behav- ioral assessments. In addition, it gives information about patient behavior in the home setting. From the pain diary one can determine behavioral patterns, defined in terms of the time of day or activity, that result in high pain versus average pain levels. It can also reveal time spent in various activities or inactivity over a week and help define the relation among pain, activity, and medication use. Iwo limitations should be borne in mind when considering pain diaries. First, their reliability is unknown and varies from person to person because diary data are dependent on the accuracy of recording. Second, whereas some people complete the form on a daily or hourly basis as directed, others do it incompletely or retrospectively just before their appointments. Furthermore, it is not known whether the experience of keeping a diary affects the experience of pain itself. Certainly, the diary calls attention to the pain and its influence on day-to-day activity. This issue awaits formal study.
220 ASSESSING kD TREATING PEN ED DYSFUNCTION The pain diary method combines the simplicity and efficiency of self- report methods with the theoretical perspective of the behaviorist. However, pain diaries presuppose that the patient (or the spouse) is a reliable and accurate historian. Ready, Sarkis, and Turner (1982) found that chronic pain patients, when asked to report medication use, gave information that was 50~0 percent below their actual drug intake. Kremer, Block, and Gaylor (1980) compared patient records with staff observations of patients' social behavior and found discrep- ancies. Sanders (1983) studied automatic monitoring of time spent out of bed ("uptime") in normal controls, psychiatric inpatients, and chronic back pain patients. He found moderate positive correlations between self-reports and automatic monitoring. All groups averaged less self-reported uptime than the automated report indicated, with the discrepancy being greatest for the chronic back pain patients. Thus, the validity of the mixed methods approach as a measure of pain is questionable. RELATED MEASURES OF DISABILITY Pain and Functional Status Of primary importance in assessment both in the context of medical care and in determining eligibility for disability benefits is ascertain- ing whether an individual is prevented from gainful employment or otherwise normal living patterns by his or her condition. The chal- lenge of assessing the patient's functional capability has attracted the attention of several investigators, and a few have attempted to relate pain to function. Disability Assessment The Northwick Park Activities of Daily Living Index has been combined with a grading system (Parish and James, 1982) to produce a way of assessing the level of functional independence of the disabled patient. Basically, the assessment procedure records whether the patient is independent or dependent on 20 different activities. There are six self-care tasks, six stages of mobility, three employment grades, and five types of domestic activity. This assessment procedure can be performed quickly and repeated as often as needed. However, it is better suited to the assessment of severely impaired individuals than to workers with pain complaints. Yvette (1980) has offered an approach to assessing the functional
MEASURING PEN ED DYSFUNCTION 221 capability of noninstitutionalized persons with polyarticular condi- tions. Activities of daily living were examined in the assessment process. Five categories (physical mobility, transfers, home chores, kitchen chores, and personal care) accounted for more than 50 percent of the variance in the data studied. Sickness Impact Persisting pain typically has a significant sickness impact on the patient. For the low back pain patient (the most-studied type of patient in the chronic pain population), this effect consists of increased hours spent reclining or in bed rest, restriction of normal social and recre- ational pursuits, emotional distress, and inability to maintain gainful employment. The concept of sickness impact is for most practical purposes interchangeable with that of disability. No comprehensive, validated objective indicators of sickness impact exist. Instead, behaviorally oriented subjective report procedures (like those discussed previously) are used. Activity diaries, in which a daily record is kept of uptime/downtime, medication use, and type of activ- ity, can be used to assess the impact of chronic pain (Follick et al., 1984~. A variety of scores can be tabulated and plotted over time from diary forms, depending on their construction. However, it is possible to go well beyond the daily diary in an attempt to systematically scale physical, social, and psychological limitations imposed by (or adopted in response to) sickness. The most fastidiously developed and fully validated of such instru- meets is the Sickness Impact Profile (SIP) (Bergner et al., 19811. This instrument is designed to be a general indicator of health status and health-related dysfunction rather than a pain-specific test. It can be a~ninistered by an interviewer or self-administered. Patients respond only to those sickness-related behavior change items that describe them appropriately. The SIP provides general scores along three dimensions of impairment physical, psychosocial, and work-recre- ation and 12 specific category scores that include, for example, communication, social interaction, and home management. Its mea- sures are derived from responses to 136 items. Follick and his colleagues (1985) investigated the SIP scores of 107 back pain patients seen at a multidisciplinary pain clinic; 75 percent of the study patients were receiving Workman's Compensation. The outcomes supported the validity of the SIP as an indicator of functional status in low back pain patients. The psychosocial dimension of the SIP was significantly correlated with the Minnesota Multiphasic Person
222 ASSESSING ED TREATING PEN ED DYSFUNCTION ality Inventory (MMPl); the physical dimension score was inversely related to independent measures of standing andlor walking and positively correlated with downtime. The investigators concluded that the SIP is a useful means of assessing functional impairment in back pain patients. Chariton, Patrick, and Peach (1983) used items from an existing health survey together with items from the SIP to produce a too] for the multivariate assessment of disability. Five global classifications for items were used: physical, psychosocial, eating, communication, and work. The disability measures derived from the items related to age and number of medical conditions but not to services utilization. Mayer anti colleagues (1986) have developed a set of rehabilitation- focused tests consisting of largely objective physical function measures for use with low back pain patients. These tests, which were combined with a battery of psychologic measures, included eight categories of measurement: (1) range of motion, (2) cardiovascular fitness and muscular endurance, (3) gait speed, (4) timed simulation of daily activities, (5) static lifting, (6) lifting under load, (7) isometric and isokinetic dynamic trunk strength, and (8) a global effort rating. These measures were obtained repeatedly through the course of treatment of back pain patients and provided information on functional capacity both to the patient and surgeon. There was an initial unemployment rate of 92 percent in the sample under study, and 82 percent of patients returned to work after treatment. Although the measures derived were not direct measures of pain, but rather measures of function, they proved to be of great value in successful treatment. Despite the successful rehabilitation of patients in this study, and with a signifi- cant decrease of self-reported pain, these patients maintained a mean VAS pain report score of 77 out of a possible 150. This suggests that rehabilitation may be achieved without major pain relief. Assessment of Work Performance In recent years, a considerable amount of attention has focused on the development of techniques to assess work-related function. Be- cause direct assessment in the workplace is usually not possible (Chaffin, 1981), various work simulation strategies have been devel- oped. The most basic approaches involve direct measures of strength or capability. For example, Harber and SooHoo (1984) used static ergo- nomic strength testing as part of a multidisciplinary evaluation program for occupational back pain. Lifting ability in several positions was quantified. Because it was independent of the degree of impair-
MEASURING PAIN AND DYSFUNCTION 223 meet, the degree of intrasubject variability could be used to detect inadequate subject efforts. This approach seems promising but falls short of true work simulation and may not be applicable to all types of occupations. The Liberty Mutual Medical Service Center in Boston uses work simulation as part of the treatment for back pain patients (Bettencourt et al., 19861. Balance monitors, pneumatic lifting and lowering equip- ment, a multi-work station, a truck driving simulator, and an upper extremity work simulator are available. The goal of the program is to allow patients to improve work performance capability and stamina while they learn to live with their symptoms. The setting has not yet been a major resource for disability assessment. This type of program is consonant with the concept of work hardening as an approach to industrial rehabilitation (Matheson et al., 19851. The adaptation of work simulation techniques to eligibility determi- nations for disability programs is intrinsically appealing because they offer promise to measure work-related function. However, a substan- tial amount of development would be required to standardize the measures and establish norms and parameters for specific kinds of jobs. Severity of Pain and Dysfunction The relation between pain severity as defined by subjective report and functional capacity has not been adequately studied. However, the literature on contingency management and operant conditioning has clearly demonstrated that the relation is not simple. Patients often fad] to perform normal daily functions because they believe, or they have been told by well-intentioned health care providers, that activity will exacerbate the pain and worsen their health. But for many chronic pain conditions the opposite is true: excessive rest and reclining contributes to the persistence of pain, and activity is beneficial rather than harmful to health. Patients who undergo a rehabilitation program involving contin- gency management typically increase their functional capacity with programmed exercise (Fordyce, 1976; Doleys et al., 1982; Roberts, 19811. Some report that the pain disappears as normal activity increases; others report that the pain remains but fades into the background of daily life. The latter type of patient demonstrates that at least some people can function in normal daily and vocational life despite the presence of pain, once they understand and prove to themselves through experience that activity is not harmful (see
224 ASSESSING ED TREATING PEN ED DYSFUNCTION Chapter 121. Still other patients are able to live reasonably normal daily lives but find that the unique demands of their vocations (typically manual labor) exacerbate the pain. In this case, vocational rehabilitation is critical to fully successful treatment. FACTORS BIASING PAIN MEASUREMENT The accurate scaling of pain in a cooperative, intelligent respondent cannot be taken for granted, and the problem of minimizing error cannot be dismissed simply because the person seems cooperative. People may knowingly or unknowingly bias their responses to any form of subjective report instrument in accordance with their beliefs, expectations, or personal goals. In addition, the test situation itself, including the attitudes and behaviors of the health care providers present, may affect the data obtained. Inaccuracy may result from distortions in memory as well as from medication toxicity. A brief description of several biasing factors follows. Personal Meaning As discussed in Chapter 8, how people interpret their symptoms and the meaning attributed to them can have a significant impact on illness behavior, including the response to treatment. Although not designed specifically for chronic pain patients, instruments that mea- sure the personal meaning of life events in terms of purposefulness, personal control, and self-esteem may be useful to adapt for pain patients. The theoretical concept of coherence as defined by Antonovsky (1980) involves the sense of put pose in life. Crumbaugh's (1968) Purpose-in-Life Scale asks respondents to indicate how they perceive their personal existences and the world as meaningful or purposeful and the extent to which their lives have reasons, purposes, or goals. A 7-point ordinal scale is used; higher scores indicate greater coherence. Personal control constructs can be assessed with instruments de- signed to scale "locus of control": the extent to which individuals believe they versus fate control the major events of their lives. The Multidimensional Health Locus of Control Scale (Waliston et al., 1978), for example, assesses the extent to which patients believe in a cause-effect relationship between their actions and their state of health. Three separate dimensions of locus of control are assessed: (1) internality, the extent to which health is perceived to be a function of one's behavior; (2) powerful others externality, the extent to which the
MEASURING PEN ED DYSFUNCTION 225 actions of other people are seen as controlling one's health; and (3) chance externality, the extent to which health or lack of health can be attributed to fate. The test consists of three six-item scales. Each item requires a response on a 4-point scale ranging from "strongly agree" to "strongly disagree." Patients' beliefs about the perceived cause, nature, and expected course and ejects of pain can be assessed with Kleinman's (1980) Explanatory Model method of elicitation. When combined with a few questions from the Health Belief Model (Rosenstock, 1966) regarding perceived vulnerability to pain, it could give assessors a quick way of taking the patient's perspective into account. Coping Chronic illness, including chronic pain, may be viewed as a stressor with which patients must cope (Lipowski, 1971; Lazarus and DeLongis, 19831. Coping skills are one more of the many factors that are likely to affect chronicity, functional status, and the seeking of disability benefits (Strauss, 1975; Thornbury, 19821. Lazarus and colleagues (Folkman, Schaefer and Lazarus, 1979; Lazarus and DeLongis, 1983) have offered a cognitive framework for understanding coping as a process. This model predicts that patients will tend to change their attitudes toward illness and associated life change demands over time. Working within this model, Thornbury (1982) and Moos and Tsu (1977) identified illness-related coping responses and several cognitive skills that may be used in therapy to teach patients to cope with extended illness. Several instruments have been developed to assess coping skills. The Ways of Coping Check List was derived from Lazarus' transactional model of stress (Folkman and Lazarus, 1980) and subsequently revised (VitaTiano et al., 19851. An event is considered stressful when the patient appraises it as being potentially dangerous to his or her psychological well-being. Five scales (revised form) can account for patient coping behaviors: (1) problem-focused coping, (2) wishful thinking, (3) avoidance, (4) seeks social support, and (5) blames self. The Coping Scale Questionnaire (Rosentie! and Keefe, 1983) asks people to rate how often they use certain strategies for coping with pain. A list of 42 strategies is provided that includes 6 different types of cognitive strategies and 1 behavioral strategy. Cognitive strategies include diversion of attention, reinterpretation of pain sensations, coping self-statements, ignoring pain, praying and hoping, and catastrophizing. The patient's belief about how much control he or she
226 ASSESSING ED TREATING PEN kD DYSFUNCTION has over pain is also recorded. Turner and CIancy (1986) used this instrument in a study of back pain patients undergoing behavioral interventions. They replicated the factor structure for the test and observed that increased praying or hoping was associated with de- creased pain ratings. Memory for Pain The ability of patients to remember pain while in a relatively pain-free state has been shown to be reasonably accurate up to a week after surgery (Hunter et al., 1979; Kwilosz et al., 1984~. For pain that persists, and when measures are being compared over time, research has demonstrated that current pain intensity produces systematic distortions of memory for prior pain, independent of treatment out- come (Etch et al., 19851. This influence of current pain on memory of past pain in persistent or chronic pain situations suggests that mea- sures derived from the patient's recollection are probably invalid and at best likely to be biased. Attributional Factors Patients, like people in general, tend to behave in accordance with their beliefs about the world and themselves. Moreover, behavior is accommodated to the social environment of the moment in psychiatri- cally normal persons. The person who believes that he or she has a serious undetected disease and who is being tested or observed by health care providers will generally act in accordance with this belief. Such a patient may unknowingly exaggerate reports of pain and related problems and will behave as he or she believes a person sick with a serious disease should. The patient who fears being labeled as a "psychogenic pain" patient may become theatrical in displaying suffering and functional impairment in the clinical setting, but when observed fortuitously in the hospital cafeteria later the same day may appear quite normal. The opposite bias may also occur. The driven, time-obsessed execu- tive who tends to deny disease may fait to report angina or other relevant symptoms as well as functional problems in an attempt to evade an illness assignment that would impair his or her personal and professional lifestyle. Thus, the patient's own beliefs (which may be shared by other family members) can severely bias the assessment of pain and pain-related impairment. Similar biases may affect treatment outcome when pain patients are
MEASURING PAIN ED DYSFUNCTION 227 subjected to surgery, drug therapy, or psychological interventions. Most of the literature in this area is concerned with the placebo effectthe tendency to report favorable outcomes in order to please the clinician, fulfill the therapeutic hopes of family or self, or achieve some other goal, such as maintaining employment. However, the opposite phenomenon also occurs; some patients fad] to report therapeutic benefit when it probably has occurred. Because both pain reports and performance on function tests are voluntary efforts, patients' claims of therapeutic failure cannot be challenged. Such behavior is largely a matter of patients acting in accordance with their own beliefs about their health and roles in society and may be consciously or uncon- sciously motivated. Research on placebo effects has helped to empha- size the importance of the context of interventions and of the doctor- patient relationship as determinants of pain report and pain behavior. Medication Toxicity Measurement error may also occur when the patient is toxic from excessive medication. Over months or years of seeking pain relief, some patients gradually develop toxicity from polypharmacy and medication overuse. The toxic patient often is irritable, has difficulty concentrating, is depressive, and may be lethargic (Hall et al., 19801. The data obtained from such patients are rarely reliable and are of little value if the purpose of data collection is to evaluate pain or pain relief. It is best to detoxify such individuals before attempting to evaluate their pain and its impact. SUMMARY Because human pain is an area of inquiry rich in complexity and multidimensionality, it is of interest to clinical and basic scientists in a wide variety of fields. It can be studied within several different conceptual models. From the viewpoint of measurement, however, this richness is also its great deficiency. The lack of consensus among investigators on precisely how pain should be defined operationally and the conceptual tensions that emerge from different interpretations of related variables such as mood disorders engender confusion and contradiction. A number of well-defined instruments are available for assessing pain and related variables. Each is bound to a theoretical position, and each has its strengths and weaknesses. None can yield unequivocal evidence of the presence of painful activity within the nervous system
228 ASSESSING kD TREATING PEN ED DYSFUNCTION apart from the patient's report and behaviors. Most experienced clinicians tend toward the eclectic: Subjective data are used to build a broad picture of the individual patient. Decisions for diagnosis and treatment are based on interpretations drawn from patterns evident among the combined measures. Because pain levels vary so much from hour to hour and day to day, ideally, such measures would be used repeatedly in order to have a longitudinal record and to avoid making inappropriate inferences based on a single assessment. In recent years measures of functional capability have emerged in the literature. Pain-related disability has been investigated in several studies, and it is clear that (1) there is no direct relation between severity of pain and disability and (2) some disabled pain patients can be rehabilitated when pain relief is not achieved. These findings suggest that disability status ought not to be granted on the basis of pain complaint alone. Evaluation of the candidate for disability status should include pain assessment, but this should be interpreted in the context of measures of functional capability in various realms of the claimant's life. IMPLICATIONS FOR SSA DISABILITY ASSESSMENT In light of the preceding review of pain measurement, the committee agrees with the conclusion of the Pain Commission (U.S. Department of Health and Human Services, 1987) that there is no direct, objective way to measure the experience of pain. Because adequate technology for the objective assessment of chronic pain is not available, the committee believes that attempts to draw inferences about the ability of a patient to engage in gainful employment on the basis of pain measurement are futile. Instead, disability evaluation efforts should focus on pain-related dysfunction rather than on pain alone. Both multidimensional pain indices and measures of functional capac- ity should be obtained so that a comprehensive evaluation is per- formed. MEASUREMENT PANELISTS Fields, Howard L., M.D., Ph.D. Professor of Neurology and Physiology, University of California, San Francisco Florence, David W., M.D. Director of Medical Affairs, People's Com- munity Hospital Authority, Wayne, Michigan Keefe, Francis, Ph.D. Associate Professor and Director, Pain Manage- ment Program, Duke University
MEASURING PAIN AND DYSFUNCTION 229 Marcus-Lewis, Frances, R.N., Ph.D. Professor of Community Health Care, University of Washington School of Nursing, Seattle Merskey, Harold, D.M., F.R.C. Psych. Director of Education and Research, London Psychiatric Hospital, Ontario Sternbach, Richard A., Ph.D. Director, Pain Treatment Center Scripps Clinical Medical Group, Inc., La JolIa, California REFERENCES Anastasi, A. Psychological testing. In: The Handbook of Clinical Psychology (Walker, C.E., ed.). Homewood, IL: Dow Jones-Irwin, 1983. Antonovsky, A. Health, Stress, and Coping. San Francisco: Jossey-Bass, 1980. Bergner, M., Bobbitt, R.A., Carter, W.B., and Gilson, B.S. The Sickness Impact Profile: development and final revision of health status measure. Medical Care 19:787-805, 1981. Bettencourt, C.M., Carlstrom, P., Brown, S.H., Lindau, K., and Long, C.M. Using work simulation to treat adults with back injuries. American Journal of Occupatiorzal Therapy 40:1~18, 1986. Carlsson, A.M. Assessment of chronic pain. I. Aspects of the reliability and validity of the visual analogue scale. Pain 16:87-101, 1983. Chaffin, D.B. Functional assessment for heavy physical labor. Occupational and Health Safety 50:24 32, 1981. Chapman, C.R., Casey, K.L., Dubber, R., Foley, Keel., Gracely, R.H., and Reading, A.E. Pain measurement: an overview. Pain 22:1~1, 1985. Charlton, J.R., Patrick, D.L., and Peach, H. Use of multivariate measures of disability in health surveys. Journal of Epidemiology and Community Health 37:296 304, 1983. Cleeland, C.S. Measurement and prevalence of pain in cancer. Seminars in Oncology Nursing 1(2):87-92, 1985 Corson, J.A., and Schneider, M.J. The Dartmouth Pain Questionnaire: an adjunct to the McGill Pain Questionnaire. Pain 19:59~9, 1984. Crumbaugh, J.C. Cross-validation of Purpose-in-Life test based on Frankl's concepts. Journal of Ind ibid ual Psychology 24:74-81, 1968. Daut, R.L., Cleeland, C.S., and Flannery, R.C. Development of the Wisconsin Brief Pain Questionnaire to assess pain in cancer and other diseases. Pain 17:197-210, 1983. Doleys, D., Crocker, M., and Patton, D. Responses of patients with chronic pain to exercise quotas. Journal of the American Physical Therapy Association 62:1111-1114, 1982. Eich, E., Reeves, J.L., Jaeger, B., and Graff-Radford, S.M. Memory for pain: relation between past and present pain intensity. Pain 23:37~380, 1985. Evans, I.M. Behavioral assessment. In: The Handbook of Clinical Psychology (Walker, C.E., ed.). Homewood, IL: Dow Jones-Irwin, 1983. Folkman, S., and Lazarus, R. An analysis of coping in the middle-aged community sample. Journal of Health and Social Behavior 21:219-239, 1980. Folkman, S., Schaefer, C., and Lazarus, R. Cognitive process as mediators of stress and coping. In: Human Stress and Cognition (Hamilton, H., and Warburton, D., eds.). New York: Wiley, 1979. Follick, M.J., Ahern, D.K., and Laser-Wolston, N. Evaluation of a daily activity diary for chronic pain patients. Pain 19:373-382, 1984.
230 ASSESSING AND TREATING PAIN AND DYSFUNCTION Follick, M.J., Smith, T.W., and Ahern, D.K. The Sickness Impact Profile: a global measure of disability in chronic low back pain. Pain 21:61-76, 1985. Fordyce, W.E. Behavioral Methods for Chronic Pain and Illness. St. Louis: C.V. Mosby, 1976. Gracely, R.H., Dubner, R., and McGrath, P. Narcotic analgesia: Fentanyl reduces the intensity but not the unpleasantness of painful tooth pulp sensations. Science 203:1261-1263, 1979. Hall, R., Stickney, S., and Earl, R. Behavioral toxicity of nonpsychiatric drugs. In: Psychiatric Presentations of Medical Illness (Hall, R., ed.). New York: Spectrum Publications, Inc., 1980. Hammonds, W., and Brena, S.F. Pain classification and vocational evaluation of chronic pain states. In: Pain Measurement and Assessment (Melzack, R., ed.). New York: Raven Press, 1983. Harber, P., and SooHoo, K. Static ergonomic strength testing in evaluating occupational back pain. Journal of Occupational Medicine 126:877~84, 1984. Hunter, M., Philips, C., and Rachman, S. Memory for pain. Pain 6:35~6, 1979. Jette, A.M. Functional capacity evaluation: an empirical approach. Archives of Physical Medicine and Rehabilitation 61:8~89, 1980. Keefe, F.J., and Block, A.R. Development of an observation method for assessing pain behavior in chronic low back pain patients. Behavior Therapy 13:363-375, 1982. Keefe, F.J., Brantley, A., Manuel, G., and Crisson, J.E. Behavioral assessment of head and neck cancer pain. Pain 23:327-336, 1985. Keefe, F.J., and Hill, R.W. An objective approach to qualifying pain behavior and gait patterns in low back pain patients. Pairz 21:15~161, 1985. Kerns, R.D., Turk, D.C., and Rudy, T.E. The West Haven-Yale Multidimensional Pain Inventory (WHYMPI). Pain 23:345 356, 1985. Kleinman, A. Patients and Healers in the Context of Culture. Berkeley: University of California Press, 1980. Kremer, E., Atkinson, J.~., and Ignelzi, R.J. Measurement of pain: patient preference does not confound pain measurement. Pain 10:241-248, 1980. Kremer, E.F., Block, A.J., and Gaylor, M.S. Behavioral approaches to treatment of chroruc pain: the inaccuracy of patient self-report measures. Archives of Physical Medicine and Rehabilitation 62:188-191, 1980. Kwilosz, D.M., Gracely, R.H., and Torgerson, W.S. Memory for post-surgical dental pain. Pain, Supplement 2, S426, 1984. Lazarus, R.S., and DeLongis, A. Psychological stress and coping in aging. American Psychologist 38:24~254, 1983. Lehmann, H.E. Affective disorders: clinical features. In: Comprehensive Teoctbook of Psychi- atry (Kaplan, H.I., and Sadock, BJ., eds.). Baltimore, MD: Williams & Wilkins, 1985. Linton, S.J. The relationship between activity and chronic back pain. Pain 21:28~294, 1985. Lipowski, A. Physical illness, the individual and the copiT~g process. International Journal of Psychiatric Medicine 1:91-101, 1971. Matheson, L.N., Ogden, L.D., Violette, K., and Schultz, K. Work hardening: occupa- tional therapy in industrial rehabilitation. American Journal of Occupational Ther- apy 39:314 321, 1985. Mayer, T.G., Gatchel, R.J., Kishino, N., Keeley, J., Mayer, H., Cupra, P., and Mooney, V. A prospective short-term study of chronic pain patients utilizing objective functional measurement. Pain 25:5348, 1986.
MEASURING PAIN AND DYSFUNCTION 231 Melzack, R. Discriminative capacity of the McGill Pain Questionnaire. [Letter to the Editor]. Pain 23:201-203, 1985. Melzack, R., and Torgerson, W.S. On the language of pain. Anesthesiology 34:50~9, 1971. Moos, R., and Tsu, V.D. The crisis of physical illness: an overview. In: Coping with Physical Illness (Mood, R., ed.). New York: Plenum, 1977. Parish, J.G., and James, D.W. A method for evaluating the level of independence during the rehabilitation of the disabled. Rheumatology Rehabilitation 21:107-114, 1982. Ready, L.B., Sarkis, E., and Turner, J.A. Self-reported vs. actual use of medications in chronic pain patients. Pain 12:28~294, 1982. Revill, S.I., Robinson, J.O., Rosen, M., and Hogg, M.I.J. The reliability of a linear analogue for evaluating pain. Anaesthes~a 31:1191 - 1198, 1976. Roberts, A.H. The behavioral treatment of pain. In: The Comprehensive Handbook of Behavioral Medicine, Vol. 2. Syndromes and Special Areas (Ferguson, J.M., and Taylor, C.B., eds.) Jamaica, NY: Spectrum, 1981. Rosenstock, I.M. Why people use health services. Milbank Memorial Fund Quarterly 44(3):9~127, 1966. Rosentiel, A.K., and Keefe, F.J. The use of coping strategies In chronic low back pain patients: relationship to patient characteristics and current adjustment. Pain 17:33~4, 1983. Sanders, S.H. Automated vs. self-help monitoring of "up-time" in chronic low back pain patients: a comparative study. Pain 15:399-405, 1983. Spielberger, C.D., Pollans, C.H., and Worden, T.J. Anxiety disorders. In: Adult Psycho- pathology and Diagnosis (Turner, S., and Hersen, M., eds.). New York: Wiley, 1984. Strauss, A.L. Chronic Illness and the Quality of Life. St. Louis: C.V. Mosby, 1975. Sy~ala, K.L., and Chapman, C.R. Measurement of clinical pain: a review and integra- t~on of research findings. In: Advances in Pain Research and Therapy, Vol. 7 (Benedett~, C., Chapman, C.R., Moricca, G., and Bonica, J., eds.). New York: Raven Press, 1984. Thornbu~y, K. Coping: implications for health practitioners. Patient Counselling and Health Ed ucation 4:~9, 1982. Turk, Die., Rudy, T.E., and Salovey, P. The McGill Pain Questionnaire reconsidered: confirming factor structure and examining appropriate uses. Pain 21:385 397, 1985. Turner, J.A., and Clancy, S. Strategies for coping with chronic low back pain: relation- ship to pain and disability. Pain 24:35~64, 1986. Tursky, B., Jamner, L.~., and Friedman, R. The pain perception profile: a psycho- physical approach to the assessment of pain report. Behavior Therapy 13:376 394, 1982. United States Department of Health and Human Services. Report of the Commission or the Evaluation of Pain. Washington, DC: U.S. Government Printing Office, 1987. Vitaliano, P., Russo, J., Carr, J.E., Marion, R., and Becker, J. The ways of coping checklist: revision and psychometric properties. Multivariate Behavioral Research 20:3-26, 1985. Wallston, K.A., Walston, B.S., and DeVellis, R. Development of the Multidimensional Health Locus of Control (MHLC) Scales. Health Education Monograph 6:107-117, 1978. Walsh, T.D. Letter to the Editor. Pain 19:9~98, 1984.