PSYCHOLOGICAL AND BEHAVIOURAL CORRELATES OFACUTE AND CHRONIC CONGRUENT AND INCONGRUENT LOW BACK PAINByHEATHER DEANNE HADJISTAVROPOULOSB.A. (Hons.), University of Saskatchewan, 1989A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF ARTSinTHE FACULTY OF GRADUATE STUDIESDepartment of PsychologyWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAApril, 1992© Heather Deanne Hadjistavropoulos, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of ^P5y( i1 6 6-jThe University of British ColumbiaVancouver, CanadaDate ^4„,,_,Le. DE-6 (2/88)ABSTRACTThe purpose of the study was to examine how thepsychological reactions to pain of acute low back pain (LBP)patients would compare to the psychological reactions to pain ofchronic LBP patients who displayed signs and symptoms which wereeither congruent or incongruent with underlying anatomy andphysiology. It was also of interest to examine whether negativecognitive and affective variables would mediate the expression ofpain in these pain groups, and whether verbal and nonverbalfacial measures of pain could be used to discriminate among paingroups.Subjects were assigned to the acute pain group if they hadrecurring or persistent pain for less than a three month period,and to the chronic pain group if they had recurring or persistentpain for longer than a three month period. The Pain Drawing(Ransford, Cairns & Mooney, 1976), the Nonorganic Physical SignsAssessment (Waddell, McCulloch, Kummell & Venner, 1980) and theInappropriate Symptom Inventory (Waddell, Main, Morris, Di Paolo& Gray, 1984) were used to assign chronic pain patients to eitherthe congruent or incongruent chronic LBP group.A physiotherapy protocol in which patients were asked togenuinely express pain, exaggerate and mask pain in response to apainful range of motion task was used to obtain a wide range offacial behaviour that would likely be relevant to understandingthe expression of pain in these patient populations. Verbalreports of pain in response to the painful range of motion task,and in response to the pain that patients experienced on a dailyi ibasis were also examined. In addition, questionnaires concernedwith coping strategies, worry and emotionality were used to tapcognitive and affective components of pain. Several demographicand patient pain related variables (e.g., medication use,disability, physical impairment) were also collected.The results of the study suggested that the acute andchronic incongruent pain patients had a similar psychologicalreaction to pain that was greater than the psychological reactionto pain of chronic congruent pain patients. In addition, theresults suggested that the acute patients and chronic incongruentpatients reported the unpleasantness of their pain to be similar,although only chronic incongruent patients differed significantlyfrom chronic congruent patients. Acute and chronic incongruentpatients were also similar on several demographic (socioeconomicstatus, compensation status) and pain related variables (regularuse of opiate analgesics, and disability). Both groups differedsignificantly on the variables from the chronic congruent paingroup. One final difference among groups was observed on ameasure of physical impairment in which it was found that thechronic incongruent patients had the greatest physical impairmentfollowed by the chronic congruent and acute pain patients. Alldifferences among groups were significant.The most important variable for optimally discriminatingamong acute and chronic pain groups was physical impairment.This, however, may have been an artifact since the physicalimpairment measure was biased toward chronic pain patientsobtaining higher scores. The most important variables foriiidiscriminating among chronic congruent and incongruent paingroups were related to the patient's negative interpretation ofpain, compensation status, and reported regular use of painkillers.In general, all subjects reported the pain they experiencedon a daily basis to be more intense and unpleasant than the painthey experienced in response to the painful leg movements. Thefacial actions that patients displayed when they were asked togenuinely express pain corresponded to facial expressions thathave been found to be associated with pain in previous studies,and included brow lowering, orbit tightening, levatorcontraction, and mouth opening. Also consistent with previousresearch were findings suggesting that, although subjects weresuccessful in masking and exaggerating pain, there were stillsome cues to deception apparent on the face.ivTABLE OF CONTENTSPage ABSTRACT^ iiTABLE OF CONTENTSLIST OF TABLES^ ixLIST OF FIGURES xiACKNOWLEDGEMENTS^ xiiINTRODUCTION 1LITERATURE REVIEW^ 5Models of Pain 5The Problem of Low Back Pain^ 7Acute versus Chronic Low Back Pain^ 8Organic and Functional Diagnoses 10Objective Assessment of Incongruent Pain^ 14Verbal Report in Pain Assessment^ 20Importance and Use of Verbal Report^ 20Limitations of Self Report^ 20Nonverbal Pain Behaviour in Pain Assessment^ 23Importance and Use of Nonverbal Pain Behaviour^23Limitations of Nonverbal Pain Behaviour^ 25Objective Measurement of Facial Behaviour 29Studies of Painful Facial Behaviour^ 31Summary and Purpose^ 37Hypotheses^ 39Cognitive and Affective Variables^ 39Verbal Report^ 40Nonverbal Facial Behaviour^ 41METHOD^ 43Subjects 43Setting^ 45Videotape Equipment^ 46Procedure^ 46Assignment of Patients to Acute and Chronic Pain Groups^50Measures Used to Assign Chronic Patients to Congruent andIncongruent Pain Groups^ 51The Nonorganic Physical Signs Assessment^ 51The Pain Drawing^ 52The Inappropriate Symptom Inventory^ 52Criteria for Assigning Chronic Patients to the Congruent andIncongruent Pain Groups^ 53Measures of Physical Impairment and Disability^ 53Physical Impairment^ 54Disability^ 55Self Report of Pain 56The Descriptor Differential Scales^ 56Self Report Measures of Cognition and Affect^ 58The Coping Strategy Questionnaire^ 58The Pain Experience Scale^ 59Videotape Segment SeleCtion 60Coding and Scoring of Facial Activity^ 62Overview of Statistical Analyses^ 65RESULTS^ 70Clarification of Chronicity of Pain Complaint^ 68Clarification of the Meaning of Incongruent Pain 68viComparison of Groups on Incongruent Pain Measures^70Further Examination of Acute Pain Patients^ 73Analyses of Experimental Conditions^ 74Analyses of Demographic Variables 76Analyses of Patient Pain Related Variables^ 78Relations Among Incongruent Pain Measures and Severity^81Choice of Covariates^ 82Analyses of Cognitive and Affective Measures^ 84Analyses of Verbal Report of Pain^ 84Analyses of Facial Activity^ 88Prediction of Pain Group 91Prediction of Membership in Chronic Pain Groups^96Predictors of Verbal and Nonverbal Measures of Pain^98Predictors of Movement Pain Intensity^ 99Predictors of Movement Unpleasantness 102Predictors of Genuine Painful Facial Activity^104DISCUSSION^ 106Overview of Discussion^ 106Clarification of Group Status^ 107Group Differences^ 112Demographic Variables^ 112Pain Related Characteristics^ 114Cognitive and Affective Measures 118Verbal Report of Pain^ 124Facial Expressions of Pain 126Discrimination Among Pain Groups^ 127Discrimination Among Chronic Pain Groups^ 129viiFacial Expressions of Pain^ 131Genuine, Masked and Exaggerated Expressions^131Relations Among Genuine Facial Actions of Pain 135Predictors of Verbal and Nonverbal Measures of Pain^137Predictors of Movement Intensity^ 138Predictors of Movement Unpleasantness 139Predictors of Genuine Painful Facial Activity^140Summary and Directions for Future Research^ 140REFERENCES^ 148APPENDICES 160A. Nonorganic Physical Signs^ 160B. Pain Drawing^ 161C. Inappropriate Symptom Index^ 163D. Consent Form^ 164E. Physiotherapy Protocol^ 165F. Debriefing Script 166G. Physical Impairment Index^ 167H. Oswestry Disability Questionnaire^ 168I. Descriptor Differential Scales 171J. Coping Strategy Questionnaire^ 173K. Pain Experience Scale^ 176L. Scoring Units for the FACS 180viiilableTable 1.Table 2.LIST OF TABLESGroup Means on Chronicity of Pain ComplaintPercentage of Incongruent Patients Meeting EachCriterionPaae6969Table 3.^Percentage of Incongruent Patients Meeting One,Two or Three of the Incongruent Criteria^71Table 4.^Group Means on Measures of Incongruent Pain^71Table 5.^Percentage of Acute Pain Patients Meeting Eachof the Incongruent Criterion^ 75Table 6.^Percentage of Acute Patients Meeting One, Twoor Three of the Incongruent Criteria^ 75Table 7.^Group Means on Categorical Demographic Variables^77Table 8.^Group Means on Age and Socioeconomic Status^77Table 9.^Group Means on Self Reported Job Satisfaction^77Table 10.^Group Means on Physical Impairment and Disability^80Table 11.^Correlations Among Measures of Incongruent Painand Severity^ 83Table 12.^Group Means and Standard Deviations on Cognitiveand Affective Variables^ 85Table 13.^Group Means and Standard Deviations on PainIntensity and Unpleasantness^ 87Table 14.^Univariate F-tests on Frequency of AUs AcrossInstructional Sets^ 89Table 15.^Mean Frequency of Occurrence of the VariablesAcross Instructional Sets^ 89Table 16.^Results of Tukey's Post Hoc Tests on Differencesin the Frequency of AUs Across Instructional Sets^90Table 17.^Results of Univariate F-tests Examining MeanIntensity of AUs Across Instructional Sets^92Table 18.^Mean Intensity of AUs Across Instructional Sets^92Table 19.^Results of Tukey's Post Hoc Tests ExaminingDifferences in the Mean Intensity of AUs AcrossInstructional Sets^ 93ixTable 20. Factor Loadings of Pain Related AUs 93Table 21. Group Means on the Discriminant Functions 95Table 22. Results of the Discriminant Function AnalysisPredicting Group Status 95Table 23. Classification Results Based on the TwoDiscriminant Functions 95Table 24. Results of the Discriminant Function AnalysisPredicting Congruent and Incongruent Pain Groups 97Table 25. Classification Results of Congruent andIncongruent Patients 97Table 26. Significant Predictors of Movement Intensity 101Table 27. Significant Predictors of Movement Unpleasantness 103Table 28. Significant Predictors of the Pain ExpressionDuring the Genuine Pain Segment 105xLIST OF FIGURESFigure^PageFigure 1.^Example of Facial Activity Observed During^181Baseline, Masked, Genuine, and ExaggeratedSegmentsx iACKNOWLEDGEMENTSMany people have contributed directly or indirectly to thiswork. I am very grateful for their assistance. First, I'd liketo thank Dr. Ken Craig for his valuable advise, comments andsupport. I'd also like to thank my research committee, Drs.Wolfgang Lindin, Jim Johnson and Darrin Lehman for their adviseand comments throughout various stages of this research.I'd also like to thank those people without whose help therewould be no study. In particular, I would like to thank thestaff at the Lansdowne Physiotherapy Clinic including thephysiotherapists, Eldon Hedden, and Peter Lamla, and the patientsecretary Rita. I am also grateful to Johanna Sale who was veryhelpful in running the study and coding data, and Christine Korolwho did all of the facial coding. Dr. Ralph Hakstain and ChrisDavis must also be thanked for their aid with the statistics.I'd like to also thank my spouse, Thomas Hadjistavropoulos,for providing balance in my life with his humor, support andcompanionship. My good friend, Theresa, who has tolerated all ofmy groaning and early morning calls also deserves specialmention.xi i1INTRODUCTIONLow back pain (LBP) is among the most frequently occurringand disabling pain conditions in Western society (Waddell, 1982).The cost of the condition in terms of medical and surgical care,financial compensation and lost work time is enormous (Mayer,Gatchel, Kishino, Keeley, Mayer, Capra & Mooney, 1986). Despitethe heavy demand that it puts on the health care system andindustry, the diagnosis and treatment of LBP remain poor (Kelsey& White, 1980).One of the most common problems in dealing with LBP is thatmany LBP sufferers have no diagnosable medical disease or injury,and traditional forms of treatment (e.g., surgery) are oftenineffective (Loeser, 1980). The degree to which treatments areineffective, however, varies between groups of acute and chronicLBP pain patients, as well as within groups of chronic LBPpatients.With acute LBP the LBP problem often resolves itself withina month, frequently without the use of medical treatment (Dilane,Fry & Kalton, 1966). With chronic LBP there is tremendousvariation in the effectiveness of treatment among sufferers. Insome cases chronic LBP may be alleviated by traditional methodsof treatment (e.g., surgery, physiotherapy), whereas in othercases these methods result in expensive medical bills and noresolution of the problem (Loeser, 1980).Recently a useful distinction has been made between chronicLBP sufferers who have pain that is either congruent orincongruent with underlying anatomy and physiology. The2congruent-incongruent pain distinction appears to account formany differences that are observed among chronic LBP sufferers.For instance, chronic LBP patients who have signs and symptomsthat are incongruent with underlying anatomy and physiology havebeen found to utilize more health care resources (Waddell et al.,1984), and to have a poorer outcome and response to surgery,rehabilitation (Doxey, Dzioba, Mitson & Lacroix, 1988; Dzioba &Doxey, 1984;• McCulloch, 1977; Taylor, Stern & Kubiszyn, 1984;Uden, Astrom & Bergenudd, 1988), and acupuncture (Lehmann,Russell & Spratt, 1983) than patients with congruent signs andsymptoms of pain. Strong correlations among variables such asineffective coping, maladaptive and dysfunctional cognitions,high levels of anxiety, and reports of high sensory activity havealso been found in patients with incongruent signs and symptomscompared to patients with congruent signs and symptoms (Reesor &Craig, 1988).Of interest in the present study was how the psychologicalreactions to pain of acute LBP patients would compare to thepsychological reactions of patients with congruent andincongruent chronic LBP. It was also of interest to examinewhether the patient groups would differ in their verbal reportand nonverbal expressions of pain. Both verbal and nonverbalbehaviour have been identified as playing an importantcommunicative function in pain assessment procedures, and werepredicted to be of use in distinguishing these different groupsof patients.3Verbal report is believed to be an essential component ofpain assessment, and many objective measures of self reportedpain have been developed (Gracely, Dubner & McGrath, 1979;Melzack, 1975). Despite the importance of verbal report inassessment, the information that is obtained through self reportmeasures is limited by many factors which may serve to amplify orattenuate the pain report (Craig & Prkachin, 1983).Due to limitations of self report, nonverbal expressivebehaviour and in particular facial expressive behaviour has beensuggested to be a useful source of information that could be usedin addition to verbal report in the assessment of pain (Craig &Prkachin, 1983). Nonverbal expressive behaviour has theadvantage of being observable, verifiable, and perhaps lesssubject to distortion than self report (Craig & Prkachin, 1983).To aid in the study of nonverbal expressive behaviour, anobjective, systematic, and atheoretical measure of facialactivity, the Facial Action Coding System (FACS, Ekman & Friesen,1978a, 1978b) has been developed.Recent investigations studying experimentally induced pain(Craig & Patrick, 1985; Patrick, Craig & Prkachin, 1986; Swalm &Craig, 1990) and clinical pain (Craig, Hyde & Patrick, 1991;LeResche & Dworkin, 1988) have shown the applicability of theFACS to the study of pain. These studies have also shown thesensitivity of FACS to variations in: 1) the severity of thepainful stimulation; 2) the pain condition studied; 3) the ageand sex of the subject; 4) the social context; and 5) subjects'attempts to fake or mask pain. Generally, the results of the4studies suggest that nonverbal expressive behaviour addsimportant and non-redundant information to that provided by selfreport (Craig, Prkachin & Grunau, In press).In the present study chronic LBP patients who wereidentified as displaying congruent or incongruent pain werecontrasted with acute LBP patients. The patients were selectedfrom a local physiotherapy clinic which likely represents a moreheterogeneous sampling of community pain patients seekingtreatment than those referred to tertiary health care settings.A physiotherapy protocol in which patients were asked togenuinely express pain, and exaggerate and mask pain was used toobtain a sample of a wide range of facial behaviour. It wasexpected that this protocol would elicit important and relevantpain expressions that would aid in the differentiation of pain inthese patient populations. The groups were compared on theirnonverbal facial expressions of pain, verbal reports of pain, andcognitive and affective reactions to pain. It was expected thatinformation obtained in this study would add to the understandingof the pain experience in acute and chronic pain patients, andwould be useful in identifying and assessing pain in congruentand incongruent LBP patients.5LITERATURE REVIEWModels of PainIn the past, pain most often has been conceptualized using asensory-specificity model of pain in which the severity of injuryor tissue damage was believed to be directly proportional to thesensation of pain (Craig, 1984; Melzack & Wall, 1988; Wall,1979). The sensation of pain was understood as a symptom of someunderlying pathology (Englebert & Vrancken, 1984), and wasbelieved to play a biologically significant role as a signal orwarning to avoid further injury (Englebert & Vrancken, 1984;Wall, 1979).This model of pain has received considerable criticismbecause it fails to account for the variability observed in theexpression of pain across individuals (Sternbach & Tursky, 1965),and within individuals from one occasion to the next (Lazarus,1986). There are many dramatic examples that serve to illustratethe lack of correspondence between tissue damage and sensation.The experience of soldiers who express minimal pain to severewounds from battle and more extreme pain in response to ineptinjections (Beecher, 1955), and amputees who experience pain inlegs that no longer exist (Melzack & Wall, 1988), serve tounderscore the fact that sensation is not always directly relatedto somatic input.Researchers, who tend to hold a more recent widely acceptedmodel of pain, view the pain experience as multidimensional innature, produced by the interaction of sensory-discriminative,motivational-affective, and cognitive-evaluative influences6(Melzack & Wall, 1988). Somatic input, previously believed to bethe sole determinant of pain, is now considered to be highlysalient, but nevertheless modified by cognitive, affective, andbehavioural components (Turk, Meichenbaum & Genest, 1983). Underthis model of pain, the immediate effects of injury, as well asthe acute and chronic pain experiences are differentiated(Englebert & Vrancken, 1984, Melzack & Wall, 1988; Wall, 1979).In the immediate phase, the sensation of pain may not evenbe experienced, and activities such as fighting and fleeing maytake precedence (Wall, 1979). In the acute phase, the sensationof pain is likely to be closely related to tissue damage,although anxiety, past experience with pain, memory for pain,cultural background, and the meaning of pain to the individualmay also contribute to the pain experience (Chapman, 1977;Melzack & Wall, 1988). In both the immediate and acute phases,pain functions to produce actions that prevent further injury, tohelp individuals learn to avoid similar future situations, and toset limitations on activity which are necessary for properrecovery (Meizack & Wall, 1988; Wall, 1979). In this way, painfunctions as a need state promoting healing, rather than as asensation signaling injury (Wall, 1979).In comparison to immediate and acute pain, chronic painunder a multidimensional model of pain is not viewed as a symptomrelated to tissue damage, but rather as a syndrome in and ofitself. It is believed to be less related to somatic input thanto psychological (cognitive, affective, and behavioural)components of pain (Meizack & Wall, 1988). Further, it is7believed to serve no useful function as either a signal ofimpending injury or as a sensation encouraging rest (Sternbach,1984). In fact, chronic pain is viewed as being detrimental tothe individual's health with dysfunctional pain behaviours (e.g.,polypharmacy, polysurgery, frequent use of health care services,limitation of work and social activities, depression andpsychological distress) characteristic of the life of the chronicpain patient (Chapman, 1977; Keefe & Gil, 1986; Sternbach, 1984,1985). In the chronic pain condition, pain persists long afterall possible healing has occurred (Chapman, 1977; Melzack & Wall,1988), and medical interventions for acute pain (e.g., bed rest,withdrawal from job and personal demands, analgesic intake) maynot only be ineffective, but may also serve to perpetuate illnessand disability (Bonica, 1980; Hrudey, 1991).The Problem of Low Back PainLow back pain (LBP) is among the most frequently occurringpain conditions in Western industrialized society. It isestimated to affect over 80% of the population at some point intheir lives (Mayer, Gatchel, Kishino, Keeley, Mayer, Capra &Mooney, 1986), and is believed to be the most frequent cause ofmorbidity, disability, and perceived threat to health in middleaged men and women (Waddell, 1982). LBP also has been found tobe associated with psychological difficulties, with LBP patientsreporting more episodes of anxiety (Frymoyer, Pope, Costanza,Rosen, Goggin & Wilder, 1980; Nagi, Riley & Newby, 1973) anddepression (Frymoyer et al., 1980) than non-LBP sufferers.8The economic cost of LBP to the health care system in termsof medical and surgical care and to industry in terms offinancial compensation and work days lost is enormous (Hrudey,1991; Mayer et al., 1986). Despite the high prevalence and costof LBP, the diagnosis and treatment of the condition remain poor(Hrudey, 1991).Acute Versus Chronic Low Back Pain Differences among acute and chronic LBP patients in theirpsychological reactions to pain have long been assumed, but notstudied empirically until recently. In general, it has beenassumed that as pain persists the patient's psychologicalreaction to pain increases. The consequence of this has beenthat cognitive and affective variables have been assumed to playa smaller role in acute LBP conditions, than in chronic LBPconditions (Sternbach, 1988). Recent investigations (Ackerman &Stevens, 1989; Philips & Grant, 1991), however, suggest that thedifferences among acute and chronic pain sufferers may not be asgreat as initially assumed, since significant statisticalYdifferences among acute and chronic LBP patients on sensory,affective, and evaluative pain scores, depression, state andtrait anxiety, negative life change scores, pain behaviour, orpain cognitions have not yet been found.Although differences among acute and chronic pain patientsin their psychological reactions to pain have not been found,differences among acute and chronic LBP in the economic costs tohealth care and industry are apparent. For instance, in terms ofhealth care utilization it is estimated that a greater percentage9(perhaps as high as 80%) of the costs of LBP problems areabsorbed by chronic pain patients (Crook, Rideout & Browne, 1984;Hrudey, 1991; Nachemson, 1982). In terms of industry, adistressingly small portion of patients with chronic pain areable to return to work (Beals & Hickman, 1972), and the longerthe period of absenteeism the less likely it is that the chronicpain patient will obtain employment (McGill, 1968).With regard to diagnosis and treatment acute and chronic LBPpatients also differ. With chronic LBP patients, it has beenestimated that as many as 80% of patients have no specificdiagnosis (Bigos & Crites, 1987), and although the condition maybe treated successfully with traditional methods of treatment(e.g., drugs, physiotherapy and surgery), more often than notthese treatments result in expensive medical bills and noresolution of the problem (Loeser, 1980). In the case of acuteLBP patients, the prognosis is significantly better, with theproblem often resolving itself within a few months withoutmedical treatment (Chapman, 1977; Dilane, Fry & Kalton, 1966).Chronic LBP has resulted in higher economic costs andgreater demands on the health care system than acute LBP, and, asa result, both clinicians and researchers have focused much oftheir attention on this condition. Their attention hasfrequently been directed toward identifying chronic LBP patientswho are unlikely to benefit from traditional forms of treatmentthat may be costly, unnecessary, and potentially harmful.10Organic and Functional Diagnoses In the past, health care professionals have frequentlyattempted to identify chronic LBP patients who were likely orunlikely to benefit from traditional modes of treatment using adichotomous diagnosis (Chapman, 1977; Main & Waddell, 1982).Patients with an obvious underlying cause for their pain havebeen classified as having an 'organic' disorder, and their painhas generally been considered real (Trief, Elliott & Stein, 1987;Turk & Rudy, 1987). In contrast, patients who have complained ofpersistent pain in the apparent absence of identifiable physicalcauses have been classified as having a 'functional' disorder,and their pain often has been considered unreal (Trief, Elliott &Stein, 1987; Turk & Rudy, 1987). The pain reports of this latergroup of patients have often been assumed to be a reflection ofpsychological problems (Biedermann, Monga, Shanks & McGhie, 1986;Chapman, 1977; Joukamaa, 1987).Support for the dichotomous distinction between organic andfunctional disorders has been variable. In one study,researchers found chronic LBP patients with nondemonstrableorganic signs to have more variable and diffuse pain descriptionsthan patients with demonstrable organic signs (Leavitt, Garron,D'Angelo & McNeill, 1979). In contrast, in other studies(Fordyce, Brena, Holcomb, De Lateur & Loeser, 1978; Swanson &Maruta, 1980) few and only marginally significant relationshipsbetween patient diagnosis and words that are chosen to describepain have been found.11Few researchers have investigated behavioural activity inpatients with varying diagnoses. In the one study where thisrelation has been investigated, however, behavioural observationsof chronic LBP patients revealed that functional patients showedless physical disability, less motivation for treatment, moreexcessive complaints, and more demands on hospital staff thanorganic patients (Donham, Mikhail & Meyers, 1984). While theobservations in this study are interesting, they are limitedsince they were not based on objective observational systems, andtherefore likely were influenced by the health care professionalsown attitudes about how pain should be expressed and tolerated.In a recent study conducted by Mahon (1991) organic LBPpatients were found to have a significantly higher pain thresholdthan functional patients. The result was attributed more to aresponse bias to report sensations as painful rather than to anincreased sensory sensitivity. That is, functional patients ascompared to organic patients had a greater bias to reportsensations as painful.In attempting to find support for the organic functionaldistinction, another set of investigations have focused ondetermining whether organic and functional patients differ inpsychological disturbance. Many studies have found that the MMPIprofiles of functional patients differ significantly from organicpatients, often in the pattern of a "Conversion V" (i.e., highscores on scales 1 (hypochondriasis) and 3 (hysteria), and lowscores on scale 2 (depression) (Donham, Mikhail & Meyers, 1984;Freeman, Calsyn & Louks, 1976; Gentry, Newman, Goldner & von12Baeyer, 1977; Hanvik, 1951; McCreary, Turner & Dawson, 1979).Many other studies, however, have failed to find such adifference or if they have, the difference has failed toadequately predict diagnosis beyond that which would be expectedby chance (Leavitt, 1985; Fordyce, Brena, Holcomb, De Lateur &Loeser, 1978; Oostdam, Duivenvoordern & Pondaag, 1981; Rosen,Frymoyer & Clements, 1980; Sternbach, Wolf, Murphy & Akeson,1973). That is, while there may be statistically significantdifferences between the MMPI profiles of organic and functionalpatients, there is a very large overlap in the profiles (Trief etal., 1987).In an attempt to explain why organic and functional patientsdo not differ in the degree of psychological disturbance,Leavitt, Garron, and Bieliauskas (1980) investigated the relationbetween psychological distress and stressful life events. Theyfound that regardless of the presence or absence of organicpathology, patients who demonstrated psychological disturbancehad a higher incidence of stressful life events.^Thedistinction between organic and functional disorders has beencriticized on the grounds that judgements of organicity areunreliable (Chapman, 1977; Turk & Rudy, 1987). For judgments oforganicity to be reliable current medical knowledge anddiagnostic procedures would have to be capable of identifying allsources of pathology (Turk & Rudy, 1987). Only in this way couldone ensure that false negatives did not arise when organicpathology was present but undetected (Chapman, 1977). A studyconducted by Rosomoff, Fishbain, Goldberg, Santana and Rosomoff13(1989) serves to illustrate this point. All of the chronic LBPpatients they studied were judged to have nonorganic physicaldisorders on the basis of a traditional physical examination.When the patients were studied more thoroughly, however,musculoskeletal disease that could be a source of nociception waspresent in all of the patients.The above findings clearly suggest that medical knowledge islimited, and as a result it is not surprising that lowreproducibility of clinical findings among experienced physiciansassessing backache is common (Nelson, Allen, Clamp & De Dombal,1979). This low inter-rater reliability can perhaps explain theabove contradictory results concerning pain description andpsychological disturbance in LBP patients judged to have anorganic or functional disorder (Sternbach, 1974).The lack of reliability of judgements of organicity is notthe only criticism of this approach, however. The usefulness andvalidity of this type of distinction have also been questioned(Sternbach, 1974). The distinction between organic andfunctional pain assumes that all symptoms are controlled by someunderlying pathogenic factor (Reuler, Girard & Nardone, 1980),that symptoms that can not be explained by pathology must have apsychological origin (Chapman, 1977; Craig, 1984; Reesor & Craig,1988), and that organic disorders and psychological problems cannot co-occur (Chapman, 1977; Turk & Rudy, 1987).This type of approach ignores the multidimensional nature ofpain or the interaction between physical, cognitive and affectivecomponents of pain to produce the final pain experience (Melzack14& Wall, 1988). Further, it ignores the fact that in studyingpain it is rare to find cases where physical variables are notinteracting with psychological variables to determine the finalpain experience (Bellissimo & Tunks, 1984).The functional-organic dichotomy has also been criticized onthe grounds that it ignores the substantial role that learningplays in the experience of pain. That is, it ignores the factthat the longer a patient has a pain problem (real or imagined)the more his or her behaviour will be governed by theconsequences in the environment, and influenced by psychologicalfactors associated with the experience of pain (Chapman, 1977;Reuler, Girard & Nardone, 1980).Objective Assessment of Incongruent PainRecently, there has been an attempt to objectify theassessment of chronic LBP. This attempt has focused ondeveloping inclusionary measures of pain that focus on the degreeto which patients display pain that is incongruent withunderlying anatomy and physiology. This approach relies on theobjective assessment of inappropriate signs and symptoms of painrather than on intuitive judgements of organicity (Reesor &Craig, 1988). That is, the emphasis is on whether there isevidence that the patient is displaying signs and symptoms ofpain that are incongruent with underlying anatomy and physiology,and not on whether the patient does or does not have an objectivephysical diagnosis (Reesor & Craig, 1988).The measures that have been developed to assess incongruentpain behaviour involve three different modes of communication:15behavioural, verbal, and pictorial. The behavioural measurefocuses on behaviours or signs which deviate from anatomicalprinciples elicited during an orthopedic examination procedure(Waddell, McCulloch, Kummell & Venner, 1980). The verbal reportmeasure of incongruent pain focuses on symptoms that areexaggerated, and do not conform to anatomy and the normal diseasecourse (Main & Waddell, 1982). The pictorial measure, referredto as the 'pain drawing', involves scoring exaggerated and non-anatomical features of drawings patients produce to illustratetheir pain (Ransford, Cairns & Mooney, 1976). The incongruentpain measures are found to be moderately correlated with oneanother (Main & Waddell, 1982; Reesor & Craig, 1988), and havethe demonstrated advantage of being reliable and easily assessedduring routine examinations (Reesor & Craig, 1988).In recent studies by Reesor and Craig (1988) and Mahon(1991) all three measures were used to differentiate betweenchronic LBP patients who displayed congruent versus incongruentpain behaviour. Chronic LBP patients were considered to displaymedically congruent illness behaviour when the behavioural signs,symptom report, and pain drawings were consistent with anatomicalprinciples. Chronic LBP patients were considered to displaymedically incongruent illness behaviour when multiple nonorganicsigns, inappropriate symptoms, or non-anatomical and exaggeratedfeatures of the pain drawing were identified.This classification system focuses on the significance andinterpretation of multiple signs as indicators of incongruentpain, and ignores the use of isolated signs or symptoms in16differentiating between normal and abnormal illness behaviour(Waddell, Pilowsky & Bond, 1990). It is important to draw thedistinction between the use of multiple indicators, rather thanisolated indicators since the use of the later has been stronglycriticized (Margoles, 1990). It is also important to emphasizethat these measures are not indexes of physical pathology as somehave come to believe (Rudy, Turk, Brena, Stieg & Brody, 1990).Rather these measures are used as an assessment of the degree towhich the patient is displaying incongruent pain behaviourindependent of the degree to which they are impaired.Using these inclusionary measures to assess incongruentchronic LBP, investigators have found chronic LBP patients withincongruent pain to use more health care resources (Waddell,Bircher, Finlayson & Main, 1984) than chronic LBP patientswithout incongruent pain. More specifically, patients withincongruent pain have been found to receive more specifictreatments (i.e., analgesics, local lumbar injections,lumboscaral supports, physiotherapy, spinal manipulation, plasterjackets, and bed rest) upon admission to a back pain clinic thanpatients with little display of incongruent pain behaviour(Waddell, Bircher, Finlayson & Main, 1984).Patients who have been identified as having incongruent painas compared to patients who have been identified as havingcongruent pain have also been found to have a poorer orthopaedicoutcome in response to surgery (Doxey, Dzioba, Mitson & Lacroix,1988; Dzioba & Doxey, 1984; Taylor, Stern & Kubiszyn, 1984; Uden,Astrom & Bergenudd, 1988). In a study by McCulloch (1977),1 7patients with few or no nonorganic signs had a 74% success rateas a result of chemonucleolysis for prolapsed interverbral disc,while patients with multiple nonorganic signs only had an 11%success rate. Incongruent as compared to congruent patients arealso found to have a poorer response to physical pain reliefmodalities such as acupuncture and transcutaneous electricalnerve stimulation (TENS) (Lehmann, Russell & Spratt, 1983).Disability or loss of function in activities of daily living(Mahon, 1991; Reesor & Craig, 1988; Waddell, Main, Morris, DiPaola & Gray, 1984; Waddell, Pilowsky & Bond, 1989) and objectivephysical impairment (Reesor & Craig, 1988; Waddell, Pilowsky &Bond, 1989) have also been found to be related to incongruentpain indicators. Further, rehabilitation, as assessed by whetherthe patient returns to work, has been found to be related towhether the patient displayed nonorganic signs (Dzioba & Doxey,1984). More than 45% of patients with few nonorganic signsreturned to work, whereas less 25% of patients with multiplenonorganic signs returned to work.Reesor and Craig (1988) investigated a possiblepsychological basis for the poor response to treatment andrehabilitation. They found strong correlations among ineffectivecoping strategies, maladaptive and dysfunctional cognitions, highlevels of anxiety and reports of high sensory activity inpatients with incongruent pain as compared to patients withcongruent pain (Reesor & Craig, 1988). Mahon (1991) alsoinvestigated this relation, and while she did not findsignificant differences between congruent and incongruent18patients, she did find a trend that suggested that patientsdiffered on cognitive appraisal and affective distress.Other investigators studying LBP also have foundpsychological variables to be important in distinguishing betweenthose with incongruent and congruent pain related behaviour. Forinstance, personality traits identified by subscales on the MMPI(Doxey, Dzioba, Mitson & Lacroix, 1988; Lehman, Russell & Spratt,1983; Waddell, McCulloch, Kummell & Venner, 1980) and depression(Main & Waddell, 1982) have been found to be more stronglyrelated to multiple nonorganic signs than to few nonorganicsigns. In studies employing the pain drawing, patients with non-anatomical and exaggerated pain drawings have also been shown tohave elevated scores on the MMPI subscales of hypochondriasis andhysteria (Ransford et al., 1976; Dzioba et al., 1984; von Baeyeret al., 1983).Interestingly, while multiple nonorganic signs related tothe MMPI clinical scales they did not relate to the F or K scaleson the MMPI (Doxey et al., 1988; Waddell et al., 1980). Thissuggests that while multiple nonorganic signs may be indicativeof patients with significant psychological impairment, they arenot indicative of patients who are malingering.Incongruent pain may also be indicative of patients whoplace an excessive focus of attention on physical functioning.Recent studies, have found incongruent pain behaviour andsymptomatology to be associated with indicators of diseaseaffirmation, and somatic preoccupation (Main & Waddell, 1982;1 9Waddell, Main, Morris, Di Paola & Gray, 1984; Waddell, Pilowsky &Bond, 1989).The objective assessment of pain that is incongruent withunderlying anatomy and physiology is a dramatic improvement overearlier attempts to assess chronic LBP patients who were unlikelyto respond favorably to treatment. The distinction appears toaccount for many differences that have been observed betweenvarying groups of chronic LBP patients. The major advantage ofthis type of assessment is that it does not conform to the notionthat injury must be directly related to pain intensity, and doesnot make any assumptions about whether the patient has an organicor functional disorder. Further, it uses various types ofinformation to gain an understanding of the pain experience(pictorial, verbal and nonverbal information is obtained),Other methods of identifying incongruent and congruent painpresentation may involve the use of the patient's verbal reportof pain and nonverbal expressive pain behaviour. Verbal andnonverbal behaviours play an important communicative function inany pain assessment procedure. Since pain is not directlyrelated to tissue damage, health care professionals must relyheavily on the patients willingness and ability to communicatetheir pain experience through these means (Fordyce, Lansky,Calsyn, Shelton, Stolov & Rock, 1984). In the followingsections, the advantages and limitations of the use of verbal andnonverbal expressive behaviour in the assessment of pain will bediscussed.20Verbal Report in Pain Assessment Importance and Use of Verbal Report. Self report hasgenerally been the preferred method of gaining an understandingof an individual's pain experience. Although this may simply bea methodological convenience, there still appear to be a numberof other reasons why clinicians and researchers rely heavily onself report. First, the presence or severity of pain is aprivate and subjective experience, and, as such, it is believedto be best understood through the words of the patient in pain(Wolff, 1978). Second, "there is no physiological measure ofpain which is either as discriminating of fine differences instimulus conditions, as reliable upon repetition or as lawfullyrelated to changed conditions, as the subject's verbal report"(Hilgard, 1969). Third, no other assessment tool allows for thereconstruction of the pain experience after the actual tissuedamage or injury has occurred (Craig & Prkachin, 1983). Finally,self report measures of pain, unlike other measures of pain,allow for the separate assessment of the multidimensionalcomponents of pain (Gracely, 1983).Limitations of Verbal Report. Despite the importance ofverbal report in pain assessment, its use does suffer fromseveral limitations (Fordyce, 1976). One such limitation inusing verbal report is related to the patient's knowledge andability to describe pain (Fordyce, 1976). In many instances(e.g, with the disabled and very young children) patients areliterally incapable of describing how they are feeling (Craig &Prkachin, 1983), and even in instances when patients do have the21skills to communicate their pain, they may still find the painand it's multidimensional qualities difficult to describe(Gracely, 1983).Pain report is also sensitive to one's educational level(Gaston-Johansson, 1984), and to one's memory for pain (Jamison,Sbrocco & Parris, 1989). More specifically, the detail of thepain report is positively related to patient education (Gaston-Johansson, 1984), and one's memory for pain tends to beinaccurate and subject to overestimation (Jamison, Sbrocco &Parris, 1989). Individual differences in anxiety, depression,attitudes, expectations, coping and response styles,psychological disturbance, and personality characteristics(Beecher, 1955; Gracely, 1983; Jacox, 1980; Kremer & Atkinson,1981; Kremer, Block & Atkinson, 1983; Teske, Daut & Cleeland,1983) can also amplify or attenuate the pain report.The report of pain is also not free from the influence ofthose who may be asking questions about pain. Factors such asthe age, sex, and perceived professional status of theinterviewer (Kremer, Block & Atkinson, 1983) have been shown tobe related to the disclosure of pain information, andcharacteristics of the pain condition itself, can influence howpain is reported as well. For example different pain syndromesare perceived to be more acceptable, and therefore are morelikely to be complained about (Hardy, 1956).The social consequences (e.g., direct or indirectreinforcement) of the pain report can also mediate the verbalexpression of pain (Fordyce, 1983). Research suggests, for22instance, that the verbal report of pain is sensitive to staffattention (Redd, 1980), and attention from significant others(Block, Kremer & Gaylor, 1980; Flor, Kerns & Turk, 1987). Inaddition, some evidence suggests that the verbal report of painis influenced by financial incentives (e.g., compensation,disability payments, and litigation awards) (Brena & Chapman,1981; Finneson, 1976). The relationship between the variables,however, is far too unclear to be conclusive (Kremer, Block &Atkinson, 1983). In some instances, financial or job incentivesmay actually result in the concealment of pain because of theconsequences that would result if the report of pain were tocontinue (e.g., athletes may minimize the seriousness of aninjury) (Craig, Hyde & Patrick, 1991).Finally, the social context may also influence the report ofpain. In a study by Craig and Weiss (1971), subjects whoreceived a series of painful electric shocks in the presence of atolerant model reported significantly less pain than subjects whoreceived the shocks in the presence of an intolerant model. In asimilar study Craig, Best, and Reith (1974) found that lowintensity shocks that were usually accepted without expressionsof discomfort were rated as progressively more painful in thepresence of a confederate who also reported them as painful. Inaddition, when the shocks were administered by an experimenterrather than through an automated procedure, the tendency to ratethe shocks as painful increased further.23Nonverbal Pain Behaviour in Pain Assessment Importance and Use of Nonverbal Pain Behaviour. Sinceinformation obtained from the verbal report of the patientsuffers from several limitations, it has been suggested thatnonverbal pain behaviours should be used in conjunction withverbal report in order to make a more accurate and completeassessment of pain (Craig & Prkachin, 1983). Nonverbal painbehaviours include behaviours such as paralinguisticvocalizations (e.g., sighs and groans), overt signs of autonomicactivity, movements of the limbs (e.g., reflexes) facialexpressions, gesticulations, and postural adjustments (Craig &Prkachin, 1983).One advantage of using overt behaviour in contrast to covertbehaviour is that it is observable and publicly verifiable(Craig, 1984; Fordyce, 1978). Further, nonverbal indices may beequally or more sensitive to some aspects of the pain experience(Craig & Prkachin, 1983), and they may also be less subject topurposeful distortion than patients' self reports (Turk & Rudy,1987). A practical reason for using nonverbal behaviours inassessment is that they likely play a more immediate role incommunicating the experience, whereas, verbal report would likelycome into play later in the sequence of events (Wall, 1979).There is much evidence to suggest that people consider andattach importance to nonverbal pain behaviours when makingjudgements of another's distress. For instance, nurses reportpreferring and relying more heavily on physiological signs, bodymovements, and facial expressions than on the patients' verbal24complaint of pain (Kahn, 1966). Further, when people makejudgements of emotional expression, they attach greatercredibility to nonverbal expressions than verbal reportespecially when the two conflict (De Paulo, Rosenthal, Eisenstat,Rogers & Finkelstein, 1978). The importance of nonverbalbehaviours is also highlighted by a study in which judgments ofothers' attempts at deception were enhanced when observers wereprovided with nonverbal as well as verbal expressive behaviour(Kraut, 1978).Further evidence that suggests that nonverbal painbehaviours could be used in the assessment of pain comes fromlaboratory studies in which naive or untrained observers usednonverbal cues in making judgments about the amount of painexperienced by others (Kleck, Vaughan, Cartwright-Smith, Vaughan,Colby & Lanzetta, 1976; Lanzetta, Cartwright-Smith & Kleck,1976). In both of these studies it was reported that untrainedjudges could reliably assess the amount of distress expressed bysubjects exposed to painful electric shocks on the basis ofobservations of nonverbal behaviour. Similarly, Prkachin, Currieand Craig (1983) and Prkachin and Craig (1985) found that naiveobservers, focusing on facial expressions alone, were able todiscriminate among volunteer subjects who received low, mediumand high electric shock intensities. In this latter study, theywere also successful in determining the amount of self-reporteddistress experienced by the subjects in response to the varyinglevels of electric shock.25Limitations of Nonverbal Pain Behaviour. The use ofnonverbal behaviour is not without its own difficulties, however.Nonverbal pain behaviours are hypothesized to be shaped by: 1)attitudes toward how one should react to and tolerate pain(Jacox, 1980; Teske, Daut & Cleeland, 1983); 2) personalitycharacteristics (Jacox, 1980) 3) response styles, states ofanxiety and depression (Beecher, 1955; Kremer, Block & Atkinson,1983; Teske, Daut & Cleeland, 1983); 4) social and culturalforces (Jacox, 1980; Teske, Daut & Cleeland, 1983); and 5)environmental contingencies such as attention, financialincentives and avoidance of responsibilities (Block, Kremer &Gaylor, 1980; Brena & Chapman, 1981; Craig, Hyde & Patrick, 1991;Finneson, 1976; Fordyce, 1976, 1978; Redd, 1980). The empiricalresearch to support many of these hypotheses, however, islacking.The largest concern with regard to the use of nonverbal painbehaviours is the degree to which they are subject to consciousor unconscious distortion. Darwin (1872) held that "in the caseof the chief expressive actions they're not learned but arepresent from the earliest days and throughout life are quitebeyond our control ..." (p. 352). Contrary to Darwin'sassertion, however, nonverbal expressions of emotional states doappear to be somewhat under our control. They are believed tocome under control through the socialization of display rulesthat govern when it is, and when it is not acceptable to expressvarious emotions (Ekman & Friesen, 1971). Socially learned and26culturally specific display rules may result in expressions thatare neutralized, masked or modified in intensity.Evidence to suggest that there are in fact culturalconventions concerning stereotypic displays of pain comes fromthe mere recognition that certain persons (e.g., actors andchildren at play) are able to and do enact them with ease (Craig,Hyde & Patrick, 1991). Experimental evidence for the existenceof cultural conventions also exists. In a study conducted byEkman (1977) subjects who viewed a stressful film in the presenceof an observer exhibited initial facial movements, and thenrapidly suppressed and replaced these movement with neutral orpositive expressions such as smiles. These facial actionsindicative of distress were not suppressed when the subjectsviewed the film alone. In a similar study conducted by Kleck etal., (1976) subjects undergoing electric shock in the presence ofan age peer observer were judged as being less expressive ordistressed than when they were alone.Another study which suggests that facial actions are underthe voluntary control of the person expressing them was conductedby Lanzetta, Cartwright-Smith, and Kleck (1976). They found thatwhen subjects were instructed to deceive they were successful inconvincing judges that they had received a more or less intenseshock than had in fact been delivered. The success one has inachieving control, however, may also be subject to influences ofthe social context. Subjects smelling pleasant and disgustingodors were less successful at hiding or expressing the experiencewhen aware that someone was in the room with them (Kraut, 1982).27That is, they leaked their evaluations more than when completelyalone (Kraut, 1982).In a study by Craig, Hyde and Patrick (1991), chronic LBPpatients were asked both to fake a pain expression in response toa nonpainful leg movement and to suppress a pain expression inresponse to a painful range of motion leg movement. A genuinepain expression in response to the painful movement was alsorecorded for the purposes of comparison. These investigatorsfound that subjects were remarkably successful in obtainingvoluntary control over their pain expressions. However, using amicroanalytic coding system of facial behaviour to study theexpressions, they found more brow lowering, and cheek raising andpulling of the lip corner folds in the faked pain expression thanin the genuine pain expression. Further, in the masked painexpression condition they found residual cues of the painexpression (e.g., marginal tension around the eyes). Theseresidual cues were interpreted in line with Ekman and Friesen's(1984) notion of micro expressions which may be part of asquelched, neutralized, or masked display of emotion. A decreasein eye blinking was also observed in both instances which wasinterpreted in line with research that suggests that reducedblinking occurs when persons are engaged in cognitive activity(Holand & Tarlow, 1972; 1975). In summary the findings reportedabove suggest that subjects do indeed have remarkable controlover their facial actions. They also suggest, however, thatleakage occurs, and that expert observers viewing facialbehaviour on videotape may well be able to detect such leakage.28While expert observers using micro analytic coding systemsmay be able to make accurate observations regarding pain, thereis some concern that general inferences drawn by observersregarding these overt behaviours may be influenced by contextualand individual difference variables, and as a result may not bequite as objective (Craig & Prkachin, 1983). That is, althoughclinicians attempt to be as objective as possible, the assessmentprocess is influenced by the clinician's own professional andpersonal experience with pain (e.g., clinicians may search foracute signs of discomfort and distress when these may not besufficient or applicable to chronic pain), the clinician'sendorsement of traditional socio-cultural beliefs about the levelof pain to be expected, and a variety of other factors such asthe clinician's personality and occupation (Bond, 1979; Dudley &Holm, 1984; Fordyce et al., 1978; Jacox, 1980; Johnson, 1977;Lenburg, Glass & Davitz, 1970; Teske, Daut & Cleeland, 1983).With regard to personality characteristics, nurturancelevels of the raters have been shown to be important indetermining observer judgments of pain in an analoguepatient/clinical relationship (von Baeyer, Johnson & McMillan,1982). Further, observer sensitivity to the nonverbal expressionof pain has been shown to influence the judgement of pain.Sensitizers are found to assign higher pain ratings thanrepressors when rating slides portraying low levels of non-verbalpain expression (von Baeyer, 1982).Variables such as instructional sets and willingness toattribute pain at varying levels of severity may influence the29judgment of nonverbal pain behaviours as well (Prkachin, Currie &Craig, 1983; Patrick et al., 1986). For instance, in a studyconducted by Prkachin, Currie, and Craig (1983) judges whobelieved that subjects were hypersensitive provided higherratings of pain than judges who did not believe that the subjectswere hypersensitive. Also, judges had more difficultydiscriminating the intensities of the shocks that subjectsreceived when the subjects had been exposed to a tolerant ratherthan a no influence control condition model (Prkachin, Currie &Craig, 1983).Characteristics of the patient being observed may also serveto influence judgements of pain. For example, when physicianswere asked to rate photographs of patients who varied inattractiveness, patients who were judged to be physicallyunattractive were rated as experiencing significantly more painthan patients who were judged to be attractive(Hadjistavropoulos, von Baeyer & Ross, 1990).Objective Measurement of Facial Behaviour. Since erroneousjudgments of pain can have detrimental effects on the patient,there has been an increased interest in the development ofsystematic, objective and sensitive measures of nonverbalbehaviour. One approach to the objective assessment of pain hasbeen to develop observational coding systems of a range of bodilyand facial behaviours (Chambers & Price, 1967; Johnson, Kirkchoff& Endress, 1975; Keefe & Block, 1982; Kendall, Williams,Pechacek, Shisslak & Herzoff, 1979; Teske, Daut & Cleeland,1983). One of the most popular measurement systems involves30recording discrete motor behaviours such as guarded movements andposture (Keefe & Block, 1982). Through the use of this codingsystem, homogeneous groups of patients who show similar painbehaviour have been identified (Keefe, Bradley & Crisson, 1990).Recently, increased attention has also been placed on theuse of facial expressive behaviour in the assessment of pain(Craig & Prkachin, 1983; Grunau, Johnston & Craig, 1990; Keefe,Bradley & Crisson, 1990; LeResche & Dworkin, 1988; Prkachin &Mercer, 1989). The amount of information the face can convey ineven a short period of time and the types of information conveyed(e.g., emotional and attitudinal) suggest that the face is likelyan important area in communication (Harper, Weins & Matarazzo,1978).The most sophisticated development to date in themeasurement of facial expressive behaviour is the Facial ActionCoding System (FACS; Ekman & Friesen, 1978a, 1979b). The majoradvantage of the FACS is that it is objective, reliable, andatheoretical. The system measures 44 separate facial actionunits (discrete movements in the forehead, eye, cheek, nose,mouth, chin, and neck regions). The action units or thecombination of the action units and the duration and intensity offacial movements can be used to describe any facial expressionnot just those presumably involved in pain. There is littleopportunity for subjective judgements since trained observers areable to use explicit definitions of the specific components offacial expression.31Studies of Painful Facial Behaviour. Many studies in bothlaboratory (Craig & Patrick, 1985; Patrick, Craig, Prkachin,1986; Swalm & Craig, 1990) and clinical settings (Craig, Hyde &Patrick, 1991; LeResche, 1982; LeResche & Dworkin, 1988; Prkachin& Mercer, 1989) have shown the applicability of the FACS to thestudy of pain. Craig and Patrick (1985) used the FACS tocharacterize female expressive behaviour in response to coldpressor pain, and to study the impact of observing models whowere either tolerant or intolerant of the shock on the expressionof pain. Narrowing of the eye aperture from below, raising theupper lip, pulling the lip corners, parting the lips or droppingthe jaw and eyes closed or blinking were all systematicallyrelated to the expression of pain. The facial actions were mostsalient at the beginning of the cold pressor task and declined inintensity over time. The nonverbal activity was inconsistentwith the report of pain that increased with exposure to the coldpressor. observing an intolerant or tolerant model had no effecton nonverbal expressive behaviour, but did have an effect onverbal report and pain tolerance. These findings suggest thatnonverbal expressive behaviour may be less susceptible to socialinfluences and may provide additional and non-redundantinformation to that provided by verbal report.The FACS has also been used to study facial expressivebehaviour in female subjects experiencing a range of painful andnonpainful shocks (Patrick, Craig & Prkachin, 1986). The effectthat observing tolerant or intolerant models had on nonverbalexpressive behaviour was also investigated in this study.Consistent with the previous study, narrowing of the eye aperturefrom below, raising the upper lips and blinking were all found tobe pain related. Unlike the above study, however, brow loweringwas also observed. The authors attributed these differences tothe different nature of the shock (brief noxious stimulus) andcold pressor (relatively more enduring and aching pain comparedto electric shock) experiences.In this study, social modeling was also found to have aninteresting effect on pain expressions. Subjects who wereexposed to the tolerant model endured shocks of greater intensityand reported them to be of the same painfulness as subjects whowere exposed to intolerant models and experienced shock of alesser intensity. Although the reports of pain of subjectsexposed to the two models were equivalent, the subjects who wereexposed to the tolerant model (and thus received higher shocks)were found to have more nonverbal pain related facial actions.These findings suggest that nonverbal pain expressions mayprovide more relevant and accurate information regarding the painexperience than verbal report.In another laboratory study (Swaim & Craig, 1990), the FACSwas used to describe both male and female expressive behaviour inresponse to painful and nonpainful shocks. Brow lowering, cheekraising, lip corner pulling, lips parting, and eyes blinking wereall found to be indicative of the pain experience. Inner andouter brow raise, cheek dimpler and tight lips, on the other handwere found to contraindicate pain. These results are quiteconsistent with those already reported.^The differences in33results that do exist, however, suggest that there are importantindividual differences in the expression of pain. In this studythe FACS was also used to examine the effects of placebos onfacial expressive behaviour. The placebo was found to influencepain expressions only in men and not in women. This findingsuggests that there may also be some important sex differences inthe expression of pain in response to placebo.In a recent study by Prkachin (1992) the consistency offacial expressions of pain across several modalities ofnociceptive stimulation (e.g., electric shock, cold pressor,pressure pain, and ischemia) was examined. The results of thestudy suggested that the bulk of the information about pain isconveyed through four facial actions or combinations of facialactions which consistently increased in likelihood, intensity,and duration across all modalities compared to a pain-freeperiod. These facial actions or combinations of facial actionswere: brow lowering, lid tightening/cheek raising, eye closing,and nose wrinkling/upper lip raising. These actions loaded onone general factor, and could be combined into a sensitive singlemeasure of pain expression. Prkachin, suggests, that the resultsprovide preliminary support for the notion of a universalexpression of pain.The applicability of the FACS to the study of clinical painhas also been demonstrated. LeResche (1982) used the FACS toidentify facial actions of slides of patients experiencing painfrom acute severe physical trauma. She found that the expressionof pain was characterized by a horizontally stretched mouth,asometimes with deepened nasolabial furrow and a lowered brow,with skin drawn tightly around closed eyes. She compared thisconstellation with the constellations of other negative emotions(e.g., disgust, fear, sadness) and found that although there issome overlap in the individual actions, the constellationsremained remarkably different.In another study of clinical pain, Prkachin and Mercer(1989) used the FACS to investigate facial activity in patientswith shoulder pathology who were undergoing active and passivearm movements and experimentally induced pain by pressure.Facial actions related to pain under these circumstancesincluded: brow lowering, narrowing and closing of the eyes, lippulling, nose wrinkling, and mouth opening. Once again, theresults indicate that facial measures of pain yield sensitiveinformation about the pain experience.LeResche and Dworkin (1988) have extended the use of theFACS to the study of facial expressions in chronictemporomandibular disorder (TMD) patients who were undergoingpainful palpations of the muscles of mastication. Facial actionsthat were observed to be indicative of pain included: tighteningthe skin around the eye, lowering the brow, closing the eyes,raising the upper lip, wrinkling the nose, and stretching thelips horizontally or opening the mouth. Generally, pain-relatedactivity reached moderate levels of intensity and occurred everyfour seconds.Using the FACS, LeResche and DwOrkin (1988) found thatsubjects expressed at least one other negative emotion (e.g.,3 " ,anger, fear, disgust, contempt, sadness) in response to thepainful palpitation procedure. Subjects who were found toexpress more than one negative emotion also were found to be moreverbally and nonverbally expressive of the pain. This findingsuggests that the expression of intense pain may be associatedwith the diffuse experience of negative affect.The FRCS has also been used to study facial expressivebehaviour in chronic LBP patients undergoing a painful range ofmotion exercise during a routine physical examination of LBP(Craig, Hyde & Patrick, 1991). Brow lowering, cheek raising,tightening of the eye lids, raising the upper lips, parting thelips and closing the eyes were all found to be associated withthe point when the patients were judged to be experiencing themost pain. As mentioned earlier, these investigators also usedthe FACS to compare genuine, faked, and suppressed painexpressions. By using the FACS, it was possible to identify somecues to deceit.Other investigators have used the FACS (LeResche, Ehrlich &Dworkin, 1990) to study facial actions (e.g., cheek raising andlip corner pull) that are associated with smiling, but aredisplayed frequently in response to the painful cold pressortask. The investigators have attempted to determine whetherthese actions were indicative of a true smile denoting happinessor an insincere smile disguising the experience of negativeemotions. In order to study this, the investigatorsdistinguished anatomically between insincere smiles and genuinesmiles and videotaped the facial expressive behaviour of female3 bchronic TMD pain patients during a baseline period, a painfulexamination and a painful cold pressor task. The greatestincrease from baseline in insincere smiles occurred in theresponse to the cold pressor task, but there was also an increasein insincere smiles during the painful palpation examination. Nosuch increase in true smiles was observed.From these studies it can be concluded that a relativelyconsistent set of facial actions is associated with theexperience of pain in response to both experimentally inducedpain (Craig & Patrick, 1985; Patrick, Craig & Prkachin, 1986;Prkachin & Mercer, 1989) and clinical pain (Craig, Hyde &Patrick, 1991; LeResche, 1982; LeResche & Dworkin, 1988; Prkachin& Mercer, 1989). In the majority of these studies (four toseven) brow lowering, tightening of the eyelids, raising of thecheeks, eyes closed or blinking, raising of the upper lip,parting of the lips or dropping of the jaw, have been identifiedas indicators of painful stimulation (Craig, Hyde & Patrick,1991). Other facial actions associated with the experience ofpain were observed to occur in only a few of the studies. Theseactions included: horizontal stretching of the mouth, pulling ofthe lip corners, wrinkling the nose, deepening of the nasolabialfurrow, and drooping eyelids (Craig, Hyde & Patrick, 1991).The variations in pain expressions that have been observedcan in part be attributed to the severity and duration of thepainful stimulus. It has been proposed that more facial actionsoccur as pain becomes more severe and enduring (Prkachin &Mercer, 1989). At low levels of pain, facial actions that are37consistently related to pain (e.g., brow lowering, and narrowingof the eye aperture) are expected to be observed. At moderatelevels of pain, middle facial actions (e.g., upper lip raise, andnose wrinkling) are expected to occur in addition to thosealready identified. Finally, with the most intense levels ofpain, lower facial actions (e.g., mouth opening, horizontalstretch mouth) are expected to come into play.Further variations in the expression of pain have beenattributed to the nature of the pain condition (i.e., low backpain, TMD, and shoulder pathology) the sex, age and individualcharacteristics of the subject, and the nature of theexperimental situation (Craig, Hyde & Patrick, 1991). Despitethe variations in pain expressions, it can be concluded thatthere is a high degree of regularity in facial expressivebehaviour associated with pain, and that the use of facialexpressive behaviour in the study of pain offers additional andnon-redundant information to that provided by self report ofpain.Summary and Purpose Until recently, acute and chronic LBP patients have beenassumed to differ in their psychological reactions to pain. Ingeneral, the assumption has been that as pain persists the morenegative the psychological reaction to pain would be (Sternbach,1985). Recent investigations, however, suggest that acute andchronic pain patients may be more similar in their psychologicalreactions to pain (e.g., they score similarly on measures ofdepression, anxiety, and pain cognitions) than has often beenassumed (Ackerman & Stevens, 1989; Philips & Grant, 1991). Theseinvestigations are important in that they represent the firstattempts to systematically and objectively compare acute andchronic pain patients. At the same time, however, they sufferfrom the fact that they failed to make a distinction amongchronic congruent and incongruent pain patients. Thisdistinction has proven to be important in that patients withchronic incongruent LBP have been found to utilize more healthcare resources (Waddell et al., 1984), have a poorer outcome andresponse to treatment (Doxey et al., 1988), more ineffectivecoping strategies, and dysfunctional pain cognitions than chroniccongruent LBP patients (Reesor & Craig, 1988). Simply comparingacute and chronic pain patients without distinguishing betweenchronic congruent and incongruent patients may have served toobscure differences that may actually exist between patients.The purpose of the present study was to compare thepsychological reactions to pain of acute pain patients, andchronic pain patients with congruent or incongruent signs. Alsoof interest in the present study was whether the patient groupswould differ in their verbal report and nonverbal expressions ofpain (masked, genuine, and exaggerated). Verbal and nonverbalbehaviour have been identified as playing an importantcommunicative function in pain assessment procedures, and thusdifferences between how the patient groups would express theirpain both verbally and nonverbally were predicted.It is of note that the patients in the present study wereall selected from a local physiotherapy clinic which likelyrepresents a more heterogeneous sampling of community painpatients seeking treatment than those referred to tertiary healthcare settings. Other studies of LBP have tended to selectpatients from this latter setting.A physiotherapy protocol in which patients were asked togenuinely express pain, and exaggerate and mask pain in responseto a painful range of motion task, was used to obtain a widerange of facial behaviour that would likely be relevant orimportant in understanding the expression of pain in thesepatient populations. Verbal reports of pain in response to thepainful range of motion task, and in response to the pain thatpatients experienced on a daily basis were also examined. Inaddition, questionnaires concerned with coping strategies, worryand emotionality were used to tap cognitive and affectivecomponents of pain.In general, the relation among nonverbal expressions ofpain, verbal reports of pain, and cognitive and affectivemeasures of pain in the different LBP patient groups wereexamined. It was hoped that information obtained in this studywould add to the understanding of the pain experience in acuteand chronic pain patients, and would also be of use inidentifying and assessing pain in congruent and incongruentchronic LBP patients.HypothesesCognitive and Affective Variables. It was predicted thatpassive coping strategies, dysfunctional cognitions, decreasedability to control and decrease pain, increased levels ofL1 0emotionality and worry would be more common in chronicincongruent LBP patients than in chronic congruent LBP patientsand acute LBP patients. In addition, the negative affective andcognitive pattern was expected to be slightly more pronounced inchronic congruent patients than acute patients, since chroniccongruent patients would have experienced their pain for a longerduration than acute pain patients. A significant sex differencewas also expected with females predicted to have higher scores onthe negative affective and cognitive variables than males.Verbal report. The cognitive and affective variables wereexpected to mediate the verbal report of pain, with the patientswho were predicted to score higher on negative cognitive andaffective variables, also expected to report more pain thanpatients who were predicted to score lower on these variables.That is, in comparison to acute LBP patients and chroniccongruent LBP patients, chronic incongruent LBP patients wereexpected to have higher scores on sensory intensity andunpleasantness of pain. These higher ratings were expected bothin response to: 1) the pain they experienced on a daily basis,and 2) the pain they experienced as a result of a leg movementthat produced discomfort during a physical examination. For allpatients, the pain experienced on a daily basis was expected tobe greater than the pain experienced in response to the painfulmovement. Chronic congruent LBP patients were expected to obtainhigher ratings on both scales than acute LBP patients, althoughthe differences were expected to be minimal in magnitude. Small41sex differences were also predicted, with females expected toreport more intense and unpleasant pain than males.Nonverbal Facial Behaviour. A distinctive pattern of facialactivity (i.e., brow lowering, cheek raising, tightening of theeyelids, raising of the upper lip, parting of the lips, closingof the eyes) was predicted to be associated with the painfulmovement for all subjects. The frequency and intensity of thesefacial actions, however, were expected to be mediated bycognitive and affective variables, with patients predicted toscore higher on the negative cognitive and affective variablesalso predicted to show more frequent and more intense painfulfacial activity than those who were predicted to score lower onthe cognitive and affective measures. That is, in comparison toacute LBP patients and chronic congruent LBP patients, the facialactions of chronic incongruent LBP patients were expected to bemore frequent and of greater intensity. Compared to acute LBPpatients, the facial actions of chronic congruent LBP patientswere also expected to be more frequent and of stronger intensity.Compared to male LBP patients, the facial actions of female LBPpatients were also expected to be more frequent and of strongerintensity. Once again, however, the differences between malesand females and acute and chronic congruent patients wereexpected to be minimal.All subjects were expected to be quite successful in maskingtheir pain expressions. However, chronic incongruent LBPpatients were expected to be less successful than acute LBPpatients, and chronic congruent LBP patients. Slight differences4 2between the latter two groups were also predicted, with acute LBPpatients predicted to be better able to mask their pain thanchronic congruent LBP patients. Male LBP patients were alsoexpected to be more successful at masking their pain than femaleLBP patients. Unsuccessful pain expressions were expected toshow more residual cues of facial activity (e.g, marginaltightening around the eyes).All subjects were also hypothesized to be quite successfulin exaggerating an expression of pain. That is, facial actionssimilar to those hypothesized to occur in response to the painfulmovement were predicted to occur when subjects were asked toexaggerate a pain expression. When subjects were asked toexaggerate, however, the facial activity was expected to be ofstronger intensity than when subjects were actually experiencingpain. Chronic incongruent LBP patients were expected to havemore success at exaggerating the pain expression than chroniccongruent LBP patients and acute LBP patients, and chroniccongruent LBP patients were expected to have more success thanacute LBP patients. Female LBP patients were also expected to bemore successful at exaggerating than male LBP patients.METHODSubjects Male and female subjects were selected from among thepatients undergoing treatment at the Lansdowne PhysiotherapyClinic in Richmond. Only those patients who were between theages of 20 and 70 years, who were experiencing pain in the lowerback at the time of the assessment, and who demonstratedsufficient command of the English language to complete thequestionnaires were used as participants in the study.Litigation/compensation claims and use of prescribed medicationswere not used as exclusion criteria from the study.From April to August of 1991 an attempt to recruit allincoming patients who met the above criteria was made. Thepatients were assigned to one of three groups of back painpatients until each group consisted of 30 subjects with an equalnumber of male and female subjects in each group. In total 104patients were approached; of these 8 refused to participate, 2were omitted from the sample because they did not meet thespecified criteria, and 4 were omitted because the group to whichthey were to be assigned already had the specified number ofpatients (15 in each group).The mean age of the sample was 44.11 years (12=14.95,range=20 to 70 years). Sixty-one percent (n=61) of the samplewere married, 12% (n=11) were single, 12% (n=11) were divorced,and 8% (n=7) were widowed. Eighty-four percent (n =76) of thesample were Caucasian anglophone, while 16% (n=14) reportedEnglish as their second language, and either East Indian (9%,4 344n=8), German (3%, n=3), Italian (1%, n=1) Hungarian (1%, n=1), orNative Indian (1%, n=1) as their first language.Socioeconomic status was rated using Blishen, Carroll, andMoore's (1987) index based on the 1981 census. The meansocioeconomic index for the sample was 36.61 (SD=13.18, range=21-72). Of the entire sample 42% (n=38) were employed, 31% (n=28)were off from work because of their injury, 18% (n=16) werehomemakers not seeking work outside of the home, and 9% (n=8)were retired. The patients who were employed or temporarily offof work rated their job satisfaction on a 10 cm visual analoguescale, and reported a mean job satisfaction of 7.30 (SD=2.57,range=0-10). The majority (67%; n=60)) of the patients were notreceiving any form of compensation, while 33% (n=30) werereceiving compensation or disability payments from the Worker'sCompensation Board, or the Insurance Corporation of BritishColumbia.The mean self-reported duration of the pain condition was4.36 years (1p=6.73, range=3 days to over 27 years). Ninepercent (n=8) of the sample had had one previous surgery, and asfew as 6% (n=5) had had as many as two surgeries. The meanpercentage physical impairment (objective loss of structuralfunctioning) as assessed by the physical impairment indexdeveloped by Waddell and Main (1984) was 7.9% (SD=8.06, range=0-29). This score is slightly below the range (10-20%) of physicalimpairment that is usually reported for LBP patients. The meanreported disability (subjective loss of functioning) was 31.38%(SD=15.87, range=4-76). According to Fairbank, Couper, Davies4 5and O'Brien (1980) a score between 20-40% suggests that patientsare moderately disabled (e.g, likely having greatest difficultiessitting, lifting, and standing, moderate difficulties withtravelling, socializing and working, and minimal difficultieswith personal care, sexual activity, and sleeping).Similar to Mahon (1991) and Yang and colleagues (1985)' medication use' was classified into the following categories:anti-inflammatory/analgesics, muscle relaxants/analgesics, andopiate analgesics. Twenty percent (n=18) of the sample reportedregularly using anti-inflammatory/analgesics, 30% (n=27) reportedregularly using muscle relaxant/analgesics, and 19% (n=17)reported regularly using opiate analgesics. On the day oftesting, however,^the use of these prescribed medications wassubstantially lower, with only 11% (n=10) reporting the use ofanti-inflammatory/analgesics, 13%muscle relaxant/analgesics, and 9%opiate analgesics.(n=12)(n=8)reporting the use ofreporting the use ofSettingSubjects were informed about the nature of the study,answered questions, and completed their questionnaires while theywere in an examination room undergoing their regularphysiotherapy treatment. They were lying slightly propped up ontheir backs and the treatment involved the application of heat ortranscutaneous electrical nerve stimulation to their backs. Thevideotaped portion of the study took place in a separateexamination room in the Lansdowne Physiotherapy Clinic. In orderto get a full view of the subject's face during the physiotherapy46protocol, the camcorder was mounted on a tripod at the end of theexamination table.Videotape Eauipment Subject's facial expressions were recorded on Fuji T-60video cassettes by using a NEC VHS movie record/playbackcamcorder system with autofocus. Following the recording of thetapes an RCA video time/date generator, model TC-1440-B was usedto provide the videotapes with a digital time display (minutes,seconds, 60ths of a second). This allowed different segments tobe selected for the coding of facial activity. For codingpurposes the tapes were played back using a Panasonic 1/2 inch,VHS video cassette recorder on a RCA model JD-975 VW 19 inch -television monitor.ProcedureLBP patients who were receiving treatment at the LansdownePhysiotherapy Clinic were asked by one of two physiotherapists ifthey were interested in participating as volunteers in the study.Eighty seven percent (n=78) of the patients were approached byone physiotherapist, and the remaining 13% (n=22) were approachedby the other physiotherapist. If the patient expressed aninterest in the study the physiotherapist arranged for theexperimenter and the patient to meet during the patient's nextscheduled appointment. The author of this study served as theexperimenter with 86% (n=77) of the patients, while a thoroughlytrained research assistant served as the experimenter with 14%(n=23) of the patients. When meeting the patient, theexperimenter would introduce herself, provide background47information, answer questions and have those who were willingread and sign a consent form (see Appendix D).The study was described as having two purposes: 1) toexamine how individuals with LBP cope, and deal with their pain;and 2) to examine how individuals respond nonverbally to adiscomforting range of motion exercise. Subjects were told thatthey would be required to complete a number of questionnairesconcerning their coping strategies and pain experience. Inaddition, subjects were told that they would be asked to follow anumber of instructions while carrying out leg movements. Theywere informed that their responses to the leg movements would bevideotaped, coded by qualified coders and perhaps viewed byjudges at some later date. Permission to obtain additionalinformation from their charts was also obtained.Patients were informed in the consent form, and throughinformation given to them prior to the study that the informationthat was collected from them was for research purposes, and wasconfidential and anonymous. Further, they were informed thattheir participation was voluntary, and that they were free towithdraw their consent to participate at any time.After the signing of the consent form, backgroundinformation on the patient was collected. This includedobtaining information regarding the patients: age, maritalstatus, occupation, first language spoken, duration of paincondition, medication consumption, compensation and litigationclaims. Subjects who were employed, or were off from workbecause of their injury were also asked to rate their job48satisfaction on a 10 cm visual analogue scale that was anchoredwith the phrases "not at all satisfied" and "completelysatisfied". All subjects were also asked to rate the intensityand unpleasantness of the pain they were experiencing on a dailybasis on two Descriptor Differential Scales (DDS; Gracely et al.,1979).Following this, information from two measures of incongruentpain were obtained. This involved asking subjects questions fromthe Inappropriate Symptom Inventory (Waddell et al., 1984), andhaving subjects complete the pain drawing (Ransford et al.,1976). The subjects were then left alone to complete a number ofquestionnaires including: 1) The Coping Strategy Questionnaire(Rosentiel & Keefe, 1983); 2) The Pain Experience Scale (Turk &Rudy, 1985); and 3) The Oswestry Disability Questionnaire(Fairbank et al., 1980)After the subject completed his or her treatment, thephysiotherapist who was treating him or her on that day obtainedinformation on measures concerned with the patient's physicalimpairment (Waddell & Main, 1984) and presence of nonorganicphysical signs (Waddell et al., 1980). The physiotherapist thendirected the patients to another examination room. The patientwas asked to put a blue gown over his or her clothes, and liedown in a supine position; the video camera was adjusted so thatonly the head and shoulders of the subjects were visible on thevideo monitor of the camcorder. Once the subject was positioned,the experimenter began video recording, and the physiotherapist49who had previously treated the patient began with theexperimental protocol (see Appendix E).The first two sets of instructions involved asking thesubject: 1) to keep a neutral expression on his of her face; and2) to provide a baseline of activity by simply wiggling his orher toes. The next three setssubjects to lift both of theirexamination table. Twenty-two percentwere unable to complete this movement,attempt to lift one leg ten inches offof instructions involved askinglegs ten inches off of the(n=20) of the patientsand were thus asked toof the examination table.All patients were able to do this. In response to this movementsubjects responded to three instructions which involved havingsubjects: 1) genuinely respond to the discomforting leg movement,2) mask their response to the leg movement, and 3) exaggeratetheir response to the movement (see Appendix E). Figure 1 showsone patient's responses to these instructions.There was a total of six possible combinations of theinstructions, so within each group the first to sixth subject,the seventh to twelfth subject and the thirteenth to fifteenthsubject were randomly assigned to one of the six possiblecombinations. With each instruction the physiotherapists readthe instructions aloud, and then give the subjects approximately10 seconds to respond before asking the subject to finish themovement. It should be noted that on average the actual lengthof time that subjects were given to respond to each instructionwas seven seconds, and not ten. This appears to be a result ofthe fact that the physiotherapists counted the seconds silently50to themselves. There were no differences in the length of timethat the subjects took to respond to the instructions, nor werethere differences between the physiotherapists in the length oftime they gave subjects to respond to the instructions. When theappropriate period of time had elapsed or the subject hadfinished the movement without instructions to do so, thephysiotherapists responded by saying "Okay, now I would like youto ...". In this way it was always possible to tell when thesubject had finished the movement despite the fact that this wasnot visible on videotape.Following this protocol subjects were asked to rate theintensity and unpleasantness of the straight leg liftingmovements on two Descriptor Differential Scales (Gravely et al.,1979). This marked the end of the experiment at which timesubjects were given the opportunity to ask questions, and wereprovided with more information concerning the study and a phonenumber where they could contact the author should they have anyfurther questions (see Appendix. F).Assignment of Patients to Acute and Chronic Pain GroupsThe subjects were assigned to either an acute LBP patientgroup, or one of two chronic LBP patient groups. The assignmentof patients to the acute and chronic pain groups followed therecommendation of the committee for the Classification of ChronicPain of the International Association for the Study of Pain(IASP). This committee suggests that "3 months [of persistent orrecurring pain] is the most convenient point of division betweenacute and chronic pain [conditions]" (IASP, 1986, p. S5).51Measures Used to Assign Chronic Pain Patients to Congruent andIncongruent Pain Groups Subjects with chronic pain were assigned to one of twochronic pain groups (an incongruent pain group or a congruentpain group) depending on the degree to which their pain complaintwas incongruent with underlying anatomy and physiology. Threedifferent assessment measures of incongruent pain were used toassign the patients to the groups. The measures included theNonorganic Physical Signs Assessment (Waddell, 1980), the PainDrawing (Ransford et. al., 1976) and the Inappropriate SymptomInventory (Waddell et al., 1984). Descriptions of these measuresare given below.The Nonorganic Physical Signs Assessment (Waddell,McCulloch, Kummell and Venner, 1980) (see Appendix A). Thismeasure was used to assess the degree to which patients displayedor reported nonorganic physical signs that did not correspond toanatomical principles. The assessment, carried out by one of twophysiotherapists, was relatively rapid since five physical signswere simply scored as present or absent. These signs included:superficial and deep non-anatomically based tenderness, report ofpain from axial loadings and simulated spinal rotation duringmock examinations, increase in straight leg raising when thepatient was distracted, disturbances of muscle strength orsensation in neighboring areas that do not correspond toneurological or anatomical substrates, and overreaction toexamination. In order to determine how reliable the ratingswere, both of the physiotherapists rated 20% of the LBP patients52on the measure. The Pearson correlation coefficient between thephysiotherapists' total scores on the measure was r=.74,indicating that adequate reliability was obtained. Otherresearchers have also found adequate inter-rater reliability(Mahon, 1991; Reesor & Craig, 1988).The Pain Drawing (Ransford, Cairns, & Mooney, 1976) (seeAppendix B). The pain drawing was used to assess the degree towhich patients magnified their pain problem pictorially (Waddell,et al., 1984). Each chronic LBP patient used a variety ofsymbols to describe the nature and quality of his or her pain onan outline of a human figure. The Ransford et al., (1976)scoring system was used by either the author of this study or afemale research assistant to quantify non-anatomical orexaggerated features of this drawing. This system involvedscoring the figure for: 1) poor anatomical localization; 2)"expansion" or "magnification" of pain; 3) specific emphases; and4) a tendency toward full body pain. Inter-rater reliabilitycalculated on 20% of the pain drawings scored by the two raterswas high (r=.83, p < .0001). High Inter-rater reliability usingthe Ransford et al., (1976) scoring system has also been foundpreviously (Reesor & Craig, 1988; Mahon, 1991).The Inappropriate Symptom Inventory (Waddell, Main, Morris,Di Paola & Gray, 1984) (see Appendix C). This measure was usedas a verbal report measure of symptoms that are vague, poorlylocalized, and generally inconsistent with known physiologicaland anatomical principles. Chronic LBP subjects were asked sevenquestions concerning whether they have experienced symptoms such53as pain, numbness or collapsing affecting the whole leg, andwhether they have had pain at the tip of their tailbone, lack ofpain free spells, intolerance to treatments or emergencyadmissions to the hospital. Positive responses to the questionswere given a score of one; they were summed together to give atotal score out of seven.Criteria for Assigning Chronic Patients to the Congruent andIncongruent Pain Groups On the basis of the criteria used by Reesor and Craig (1988)and Mahon (1991) chronic LBP patients were assigned to theincongruent pain group if they obtained: 1) a score of two orgreater on the nonorganic physical signs measure; 2) a score offive or greater on the pain drawing; or 3) a score of three orgreater on the inappropriate symptom inventory. In the absenceof these criteria, chronic pain subjects were assigned to thecongruent chronic LBP group. The incongruent pain measures weredeveloped for use with chronic pain patients, and as a resultthey were not used to further subdivide the acute pain patients.Measures of Physical Impairment and DisabilityMeasures of physical impairment and disability were obtainedon all subjects to determine the severity of the back painconditions (Waddell & Main, 1984). "Physical impairment is 'ananatomical or pathological abnormality leading to loss of normalbody ability', whereas disability is 'diminished capacity foreveryday activities' or 'the limitation of a patient'sperformance compared with a fit person of the same age and sex"(Waddell & Main, 1984, p. 204).54Physical Impairment. Physical impairment was measured byone of two physiotherapists using the Physical Impairment Index(PII) developed by Waddell and Main (1984) (see Appendix G). Itis a standard set of reliable indices of organic impairment andphysical limitation. In calculating the PII, consideration istaken of the anatomic pain pattern, time pattern of attacks,previous spinal fractures and lumbar surgery, nerve compression,lumbar flexion and straight leg raising in both legs. Thenumerical loadings of these physical characteristics are given inAppendix G.It is of note that this is not an ideal measure of physicalimpairment in acute LBP patients. This is related to the factthat two of the items on the physical impairment inventory (i.e.,time pattern of attacks, and surgery) are such that chronic painpatients will always likely obtain higher scores than acute painpatients. Because there were no other measures that would allowthe physiotherapists to assess physical impairment in a shortperiod of time within the physiotherapy setting, it was decidedto use this index of physical impairment in the present studydespite its short coming. The limitations of the findings withrespect to this measure are noted.The scale yields a percentage in which 0% impairment isindicative of no impairment, 40% is indicative of the worstpossible back problem, and 100% is indicative of total bodyimpairment, which would be present only in cases of totalparaplegia or the amputation of both legs. Waddell (1987)reports that, in practice, the scale gives a conservative55estimate with normal people scoring 2.6% and LBP patients scoringon average 10-20%, and only occasional patients with LBP scoringover 30%. The total PII inter-rater reliability calculated byhaving two physiotherapists score the same 20% of back patientswas found to be very high (r=.88). This is consistent withWaddell (1987) who also has found high inter-rater reliability.Disability. The Oswestry Low Back Pain DisabilityQuestionnaire (OLDQ; Fairbank, Couper, Davies & O'Brien, 1980)(see Appendix H) was used to measure subjective disability. Thequestionnaire consists of 10 questions reflecting various degreesof pain related disability in problem areas such as analgesicmedication consumption, personal care (e.g., washing anddressing), lifting, walking, sitting, standing, sleeping, sexualand social activity and traveling.Each question contains 6 statements with each successivestatement describing a greater degree or difficulty than thepreceding statement. Each question is scored on a 0-5 scale,with 5 representing the greatest disability. The scores for eachquestion are added together, giving a total possible score of 50.The total is then doubled and expressed as a percentage.Disability is easily interpreted: 0-20% indicates minimaldisability; 20-40% indicates moderate disability; 40-60%indicates severe disability; 60-80% is indicative of those whoare crippled; and 80-100% is indicative of those who are eitherbed bound or of those who are exaggerating.The scale's validity has been demonstrated by 'airbank etal., (1980). They found that LBP patients with a strong56likelihood of spontaneous recovery showed marked improvement intheir disability scores over two and three week intervals. Theyalso found the questionnaire to have high internal consistencyand test-retest reliability.Self Report of PainThe Descriptor Differential Scales (DDS). This measure,consisting of two 13 item verbal descriptor scales (Gracely,Dubner & McGrath, 1979) (see Appendix I), was used to measure thesensory intensity and unpleasantness of the pain the patientswere experiencing on a daily basis, as well the pain the patientswere experiencing during painful range of motion tasks. Subjectswere asked to pick one set of words from each column. Althoughthe verbal descriptors of pain are not physically measurable,cross modality matching methods have been used to scale andverify the relative magnitude dimensions implied by thedescriptors (Gracely, Dubner & McGrath, 1979; Gracely, McGrath &Dubner 1978a). One advantage of using ratio scales produced bycross modality matching techniques is that they are lesssensitive to certain response biases that are prominent incategorical rating scales (Gracely, 1979, 1983; Gracely & Dubner,1981). Further, the ratio scales of verbal descriptors allow formeaningful statements about pain magnitudes (Gracely, 1979).Reliability coefficients of a group of eight subjects in twosessions one week apart, and reliability coefficients betweengroups of similar subjects have been reported as 0.96 and 0.89for sensory intensity and unpleasantness respectively in bothinstances (Gracely, McGrath & Dubner, 1978a). The scales are57seen to be particularly objective because there is substantialagreement between individuals in the values they attach to eachadjective (Gracely, 1979, 1983; Gracely & Dubner, 1981). Incomparison to visual analogue ratings of sensory intensity andunpleasantness, the DDS is as effective in quantifying sensoryintensity and affective aspects of pain, and superior indifferentiating or separating the two from each other (Duncan,Bushnell & Lavigne, 1989).The validity of distinguishing between sensory intensity andunpleasantness has been demonstrated by studies that show thescales are differentially sensitive to placebo, narcotic andtranquilizing drugs (Gracely, Dubner & McGrath, 1979; Gracely,McGrath & Dubner, 1978b). Specifically, fentanyl, a short actingnarcotic, reduced the sensory intensity, but not theunpleasantness of electrically induced tooth pulp sensations. Incomparison to this, a saline placebo reduced the unpleasantnessbut not the sensory intensity of the sensations (Gracely, Dubner& McGrath, 1979). In another study, diazepam, a minortranquilizer, altered the unpleasantness but not the sensoryintensity of electrocutaneous stimuli (Gracely, McGrath & Dubner,1978b). These studies serve to demonstrate that it is not onlyvalid to distinguish between sensory intensity andunpleasantness, but that if the pain experience is to be studiedcritically, these two dimensions of pain must be examinedindependently.In general the scales are still considered to be reliableand advantageous because they: 1) differentiate between varying58dimensions of pain: sensory intensity and unpleasantness; 2)anchor subjective responses allowing for within-subject andbetween-subject comparisons; and 3) can be used in describingboth naturally occurring acute and chronic pain andexperimentally induced pain (Gracely, 1979, 1983; Gracely &Dubner, 1981).Self Report Measures of Cognition and Affect The Coping Strategy Ouestionnaire (CSQ). The CSQ (Rosentiel& Keefe, 1983) (see Appendix J) was used to assess the frequencythat patients report using one of seven strategies to cope withpain. These strategies include: diverting attention,reinterpreting the pain sensations, employing coping self-statements, ignoring the pain sensations, praying or hoping,catastrophizing and increasing behavioural activities. Eachcoping strategy subtype is represented by six items on thequestionnaire. Using a seven-point scale (0=never and 6=always),subjects indicated how often they used each item when theyexperienced pain. In addition, the CSQ includes two measures ofoverall effectiveness of coping strategies. These measuresinvolved having subjects rate their ability to control pain andtheir ability to decrease pain on a seven point scale where 0=nocontrol/can not decrease pain, and 6=complete control/candecrease pain completely.The factor structure of the Coping Strategy Questionnairehas been examined on five different chronic pain populations(Lawson, Reesor, Keefe & Turner, 1990). From this examination athree factor structure has been suggested. The first factor59reflects a conscious active use of cognitive strategies or copingprocesses, and is represented by three scales concerned withignoring pain sensations, employing coping self statements, andreinterpreting pain sensations. The second factor reflects aself-evaluative component that is represented by two questionsconcerned with the patient's ability to control pain, and thepatient's ability to decrease pain. The third factor reflectspassive strategies for dealing with pain, and is represented bythe scales concerned with praying and hoping and divertingattention. Two scales that are covered by the CSQ do notconsistently relate to any of the three factors. These scalesare treated as a separate measures reflecting behavioural copingstrategies and tendency to employ catastrophizing cognitions.In general, the pain-coping strategies are found to behighly predictive of psychological distress and pain report inchronic pain patients (Keefe, Crisson, Urban & Williams, 1990).The catastrophizing scale is a particularly good measure ofemotional and behavioural adjustment to pain in chronic painconditions (Reesor & Craig, 1988; Rosentiel & Keefe, 1983; Turner& Clancy, 1986). Rosentiel and Keefe (1983) and Turner andClancy (1986) have found a high inter-item correlation (alphacoef. = 0.71 - 0.85), indicating that the test is internallyreliable.The Pain Experience Scale (PES). The PES (Turk & Rudy,1985) (see Appendix K) was used to assess the subject'scognitive-evaluative reaction to chronic pain. Six items in thequestionnaire reflect how worried the subject is, and 13 items60reflect how emotional the subject is. For all questions,subjects reported how frequently they felt certain emotions orthought certain thoughts on a seven-point scale, where 0 equalednever and 6 equaled very often. The average of the 6 itemsreflecting worry, and the average of the 13 items reflectingemotionality was calculated.Factor analysis has revealed that the two scales arereliable with alpha=0.91, p < .001 for the emotionality scale,and alpha=0.74, P < .001 for the worry scale. The test-retestreliability after a two week interval is high (r=0.89, p<.001 andr=0.81, 2<.001 for each scale respectively). The scales are alsosensitive to cognitive behavioural treatment, with significantpre-post changes on both scales being observed followingtreatment of 30 pain patients.Videotape Segment SelectionFive 4-second segments from each subject's videotape wereselected for coding. The first of these was a "neutral" segment,in which the subject was at rest and his or her face wasexpressionless. The segment was taken from the beginning of thesession when the subject was asked to keep a neutral expressionon his or her face. The four seconds prior to thephysiotherapist's indication to the patient that they no longerhad to keep a neutral expression was used as the segment. Thissegment served as a reference for the coders when they werecoding other segments; it provided information about individual'sfacial structure, lines, wrinkles, etc. which might otherwise61influence the coder's judgments of the presence or absence ofparticular action units.The second segment was a "baseline" segment that was takenfrom the period when subjects were asked to wiggle their toes.Since it was difficult to determine when the subject beganwiggling his or her toes, the physiotherapist said "okay" oncethe subject had finished the movement. In that way the 4 secondsprior to the physiotherapist's verbalization were used asrepresentative of the activity during the baseline segment.The third segment was taken from the period when subjectswere asked to genuinely express their pain. Two secondspreceding and two seconds succeeding the point that theexperimenter judged the subject to be expressing the greatestamount of distress were used. If no movement was visible to theexperimenter, then four seconds prior to the physiotherapist'srequest for the patient to complete the movement was taken as thegenuine pain segment. In order to calculate the reliability ofthis judgement, a second coder rated 20% of the patients. ThePearson correlation coefficient between the judgements was veryhigh (r=.99, 2 < .0001), and the average difference between thesegments selected by the experimenter and the research assistantwas only 0.48 seconds.The fourth segment corresponded to the "masked" expressionof the subject. Since the subject was masking his or her facialexpressions, it was difficult to point out when he or she wasexperiencing the most pain. Therefore, the physiotherapist wasasked to say "okay" once the movement was completed. Once again,62the 4 seconds prior to the physiotherapist's verbalization wasused as representative of the activity during the masked segment.The fifth segment corresponded to the period when thesubject was "exaggerating" his or her pain. Two secondspreceding and succeeding the moment that the experimenter judgedthe greatest amount of activity to be occurring was used for thissegment. If no observable facial activity was present, the fourseconds prior to the physiotherapist's request for the patient tocomplete the movement were taken as representative of the"exaggerated pain segment". In order to determine thereliability of the judgement, a female research assistant rated20% of the patients. The Pearson correlation coefficient betweenthe judgements was very high (r=.99, R < .0001), and the averagedifference between the segments selected by the experimenter andthe research assistant was only 0.13 seconds.Coding and Scoring of Facial ActivityEach segment (i.e., 2-5) of the videotape was scored for all44 facial action units (AUs) specified by the FACS. However, twosets of actions were combined to make two new variables. AUs 6(cheek raise) and 7 (lid tighten) were combined into one variablerepresenting orbit tightening. There is a precedent forcombining these variables because the forms and muscular bases ofthe movements are similar (Prkachin, 1991). AUs 9 (Nose Wrinkle)and 10 (Upper Lip Raise) were also combined into one variablerepresenting levator contraction. There is also a precedent fordoing this because the movements involve contractions of the samemuscle (i.e., the levator labili) and are believed to be63different stages of the same expression (Prkachin, 1991; Prakchin& Mercer, 1989). A complete list of the facial AUs is listed inAppendix L.For each AU scored in a particular segment, the frequency ornumber of occurrences of each AU was recorded. Further, for themajority of the AUs a standardized five-point rating scale thatranged from A (trace) to E (maximum) (Ekman & Friesen, 1983) wasused to code the AUs for intensity.^The average intensity ofeach AU in each segment was then calculated.There were a few exceptions to the above scoring method.First, AUs 11 (Nasolabial Deepen) and 38 (Nostril Dialation) havenot lent themselves to intensity coding (Ekman & Friesen, 1983;Prkachin, 1991) and were thus simply coded as present or absent.Second, anytime AUs involved the opening of the mouth (e.g., AU25 lips apart, AU 26 jaw drop, and AU 27 mouth stretch) a newvariable called "mouth opening" was coded as present. It wasthen given an intensity score of '1' if the lips were parted (AU25), an intensity score of '2' if the jaw had dropped (AU 26) andan intensity score of '3' if the mouth was stretched open (AU27). This is the same procedure that has been used by Prkachinand Mercer (1989).While the FACS data coders coded segments the volume on thetelevision was turned down so that they were blind to the natureof each segment (other than the neutral segment). The primarycoder was also completely blind to the group membership of eachsubject. The second coder was the author of the study, and as aresult was familiar with all of the subjects prior to coding.64However, she did not specifically remember the patient's groupmembership since the data had been collected anywhere from one tofive months prior to her coding. The coders were both thoroughlytrained and experienced in the use of the FACS, and hadsuccessfully met the reliability criteria for scoring required(Ekman & Friesen, 1978b) for certification as proficient FACScoders.The primary coder coded 97% (n=87) of the subjects. Thesecond coder provided primary coding for 3% (n=3) of thepatients, and in addition provided inter-rater reliability codingon an additional 20% (n=18) of the patients. Percent agreementwas calculated according to the formula recommended by Ekman andFriesen (1978a):No. of Agreements x 2Percent Agreement =Total no. of items scoredAn agreement was scored if the coders agreed on the occurrenceof the AUs within a segment. This form of percent agreement ispreferred over one that utilizes both occurrence andnonoccurrence agreement. If the latter is used, the reliabilitytends to be inflated, since the nonoccurrence of AUs tends to befar more frequent than the occurrence (House, House & Campbell,1981). This formula yielded a percent agreement of 84%. Inorder to calculate reliability of judgements of intensity of AUs,the intensity scoring for each agreed upon AU of the two coderswere correlated. The Pearson correlation coefficient wasmoderately high (r=.80).65Overview of Statistical Analyses Nine different sets of analyses were carried out. The firstof these served to clarify the meaning of acute, chroniccongruent and chronic incongruent pain. An ANOVA was used toexamine whether the groups differed in chronicity of paincomplaint, and a MANOVA was used to examine group and sexdifferences on the incongruent pain measures. In addition,several Chi Square analyses were carried out to examine how manysubjects within the acute and chronic incongruent pain groups meteach of the criteria.A second set of analyses examined whether there were groupor sex differences in the implementation of experimentalconditions. The third set examined group and sex differences ondemographic related variables, and the fourth set examined groupand sex differences on pain related background variables. Inthese analyses categorical variables were analyzed using ChiSquare statistics with group and sex as the independentvariables. When differences were found Marascuilo's (1980)procedure for carrying out multiple comparisons was used.Continuous variables within each set of the above analyses wereanalyzed using MANOVA with group and sex as the independentvariables. When the multivariate tests were significant,univariate F tests were examined and followed by Tukey's post hoctests when appropriate.In a fifth set of analyses a MANOVA examined group and sexdifferences with respect to the cognitive and affectivevariables. In a sixth set of analyses a repeated measures MANOVA66was used to examine group and sex differences on reported painintensity and unpleasantness either on an average day or inresponse to the leg movements. In the above analyses when themultivariate tests were significant, the univariate F tests wereexamined and followed by Tukey's post hoc tests when appropriate.A seventh set of analyses was used to examine the facialexpressions of pain. This involved carrying out two repeatedmeasures MANOVA's to examine group, sex and repeated measures(e.g., baseline, masked, genuine, and exaggerated) differences inthe frequency and intensity of a select group of AUs (i.e., thoseAUs which occurred more than five percent of the time, or werepreviously found to be pain related). The frequency scores forthose AUs which were found to occur more frequently during thegenuine segment compared to the baseline segment were thenentered into a principal components analysis to see how theaction units inter-related.In the eigth set of analyses the variables which were foundin the above analyses to differ among groups were entered into adiscriminant function analysis. This allowed us to determinewhich of the variables optimally discriminated among the paingroups, and was particularly important since it allowed us todetermine the relative importance of demographic, and patientpain related variables compared to the dependent variables ofinterest. A second discriminant function analysis was carriedout, but this time only chronic congruent and incongruent groupswere examined. This was useful in that the results could be67directly compared to previous research that did not examine acutepain in relation to chronic pain.A final set of analyses were carried out to examine which ofall the variables measured in the study related to the verbal andnonverbal measures of pain.^The three criterion variables ofinterest were: 1) the reported intensity of the painfulphysiotherapy movement; 2) the reported unpleasantness of thepainful physiotherapy movement; and 3) the summed painful facialactivity score for the genuine pain segment. Six groups ofpredictor variables were examined for their relation to each ofthe criterion variables of interest: 1) group assignmentmeasures; 2) demographic background variables; 3) patient painrelated variables; 4) cognitive and affective variables; 5)verbal report variables; 6) and summed painful facial activityscores for each instructional set. For each criterion variable afinal stepwise multiple regression analysis was carried out inwhich those variables which were found to be significantlyrelated to the criterion of interest served as the predictorvariables. The above procedure, although somewhat complicated,served to keep the ratio of number of subjects to predictorvariables at a minimum.68RESULTSClarification of Chronicity of Pain Complaint An ANOVA with group and sex as the independent variables,and chronicity of pain complaint as the dependent variable wasused to examine: 1) whether the separation of patients intodistinct acute and chronic pain groups was successful; and 2)whether there were any differences between male and femalepatients. This analysis revealed a significant main effect forgroup (F (2, 84)=12.76, p < .0001), but not for sex. Nointeractions were found. Tukey's post hoc comparisons revealedthat the chronic pain groups did not differ significantly fromeach other on chronicity of pain complaint, but did differsignificantly from the acute pain group. On average chronic painpatients had an average pain complaint of 6.5 years, whereasacute pain patients had an average pain complaint of less thanone month (29 days). Table 1 presents the mean chronicity ofpain complaint in years for the groups.Clarification of the Meaning of Incongruent PainSince patients had to meet only one of the criteria from oneof the measures to be assigned to the incongruent pain group itwas of interest to examine how many patients within theincongruent pain group met each of the criteria. Table 2 liststhe percentage of male and female patients who met each criterionwithin the incongruent LBP group. Of note when examining Table 2is the relatively few patients who met the criterion for the paindrawing compared to the number of patients who met the criterionfor the nonorganic physical signs assessment, and theTable 1Group Means on Chronicity of Pain ComplaintVariable^ GroupAcute^Congruent^IncongruentM^F M^F M^FChronicity^0.10 0.06^4.64 5.70^6.86^8.81(years)n=15Table 2Percentage of Incongruent Patients (N=30) Meeting Each CriterionSex NonorganicSigns (>1)Inappropriate^Pain DrawingSymptoms (>2)^Score (>4)Males^40 ( 6)^73 (11)^27 (4)Females^80 (12)^47 ( 7) 27 (4)n=15697 0inappropriate symptom inventory. Chi square analyses wereperformed to examine whether there were sex differences inmeeting the criteria. There were no differences between thenumber of males and females who met the criterion on any of themeasures.The congruency among the measures in identifying incongruentpain patients was examined by comparing the number of patientswho met one, two or three of the criteria. Table 3 representsthe percentage of male and female pain patients in the medicallyincongruent group who satisfied one, two or three of thecriteria. As can be seen in the table only one patient met allthree criteria, and about an equal number of men and women metone or two of the criteria. Of those identified as incongruenton the basis of only one of the three criteria, seven wereclassified as incongruent on the basis of the inappropriatesymptom inventory, nine were classified as incongruent on thebasis of the nonorganic signs assessment, and only one wasclassified as incongruent on the basis of the pain drawing.Comparison of Groups on Incongruent Pain Measures To clarify the differences that were created by assigningpatients to the various pain groups, a MANOVA with group and sexas the independent variables, and the three incongruent painmeasures as the dependent variables was carried out. Table 4presents the mean scores for each of the groups on each of theincongruent pain measures. This analysis revealed a significantmain effect for group, F (6,166)=9.79, p < .0001. Univariate F-tests revealed that there was a significant main effect for groupTable 3Percentage of Incongruent Patients Meeting One. Two or three ofthe Incongruent CriteriaNumber of CriteriaSex One Two ThreeMales 60^(9) 40^(6) 0^(0)Females 53^(8) 40^(6) 6^(1)n=15Table 4Group Means on Measures of Incongruent PainVariableAcuteM^FGroupCongruentM^FIncongruentM^FNonorganic 1.13 0.93 0.13 0.20 1.00 2.93SignsInappropriate 1.13 1.27 0.60 0.93 3.73 2.60SymptomsPain Drawing 1.27 1.40 0.93 0.93 3.20 3.07n=157172on each of the incongruent pain measures including the nonorganicphysical signs assessment (F (2,84)=17.96, 2 < .0001), theinappropriate symptom inventory (F (2,84)=28.68, p < .0001), andthe pain drawing (F (2,84)=6.34, p < .01).Tukey's multiple comparisons of the pain drawing and theinappropriate symptom inventory revealed that the chronicincongruent patients had significantly higher scores on thesemeasures than chronic congruent patients, and acute painpatients. The latter two groups did not differ significantlyfrom each other on either of the measures. Multiple comparisonsof the differences among groups on the nonorganic physical signsassessment revealed that the chronic incongruent patientsobtained a significantly higher score on this measure than theother two groups. In addition, the acute pain group obtained asignificantly higher score on this measure than the chroniccongruent pain group.The MANOVA of the incongruent pain measures with group andsex as the independent variables also revealed a significantgroup by sex interaction, F (6,166)=3.62, p < .002. Univariate Ftests of the interaction revealed that there was only asignificant interaction on the nonorganic physical signsassessment measure, F (2,84)=20.27, 2 < .001. Simple effectsanalysis of the nonorganic physical signs assessment revealedthat among males there were significant group differences, E(2,84)=3.28, p < .05. Tukey's post hoc comparisons of thisdifference revealed that the males in the chronic incongruentgroup and the acute pain group did not differ from each other,7 3but did differ from the chronic congruent group. Simple effectsanalysis also revealed that there were significant groupdifferences among the females, F (2,84)=22.24, 2 < .001. Tukey'spost hoc comparisons examining the differences among femalesrevealed that the chronic incongruent pain group obtained thehighest scores on this measure followed by the acute pain groupand then the chronic congruent pain group. All differences amonggroups were significant.Simple effects analyses were also used to examine whethermales and females differed within each group on the nonorganicsigns measure. Here it was found that the males and femalesdiffered from each other only within the chronic incongruent paingroup, F (1,84)=20.76, 2 < .001. In this group it was found thatfemales obtained significantly higher scores than males on thenonorganic signs measure. Upon further examination of thismeasure it was found that females as compared to males were morefrequently observed to overreact nonverbally to the physicalexamination, (X 2 (1)=8.56, 2 < .01), and to report pain inresponse to axial loading (X 2 (1)=10.16, p < .002). Morespecifically, 80% of the incongruent females as compared to 27%of the incongruent males showed exaggerated facial expressions ofpain during the examination, and 60% of the incongruent females,compared to 0% of the incongruent males reported pain in responseto the axial loading.Further Examination of Acute Pain Patients From the above results it appears that the acute painpatients were more similar to chronic congruent patients than74incongruent pain patients in the scores they obtained on theincongruent pain measures. However, upon further examination oftheir scores on these measures it is clear that there is widevariability in the scores. Table 5 shows the percentage ofpatients within the acute pain group who met the criteria for theincongruent pain measures. Table 6 shows the number of patientswho met one, two or three of the criteria. Together theseresults suggest that it may in the future be useful to examinedifferences among congruent and incongruent acute pain patients.Unfortunately, the small number of subjects prevented such ananalysis in the present study.Analyses of Experimental Conditions Chi square analyses were carried out to ensure that therewere no differences among groups in the experimental conditions.There were no group or sex differences with regard to whichphysiotherapist, or experimenter carried out the study or withregard to the order in which the instructions to either genuinelyexpress pain, mask or exaggerate pain were presented. The numberof subjects within each group who lifted one leg instead of twooff of the examination table did not differ among groups, but diddiffer between sexes X 2 (1)=7.78, p < .005. There were morefemales (n=16) unable to lift both of their legs off of theexamination table than males (n=4). A repeated measures MANOVA(with instructions to genuinely express, mask or exaggerate painas the repeated measure) with group and sex as the independentvariables was used to determine whether there were group, sex orrepeated measures differences in the number of seconds the75Table 5Percentage of Acute Pain Patients Meeting Each of the IncongruentCriterionSex^Nonorganic^Inappropriate^Pain DrawingSigns (>1) Symptoms (>2) Score (>4)Males^33 ( 5)^20 ( 3)^13 ( 2)Females^33 ( 5)^20 ( 3) 6 ( 1)n=15Table 6Percentage of Acute Patients Meeting One, Two or Three of theIncongruent CriteriaNumber of CriteriaSex^ One^Two^ThreeMales^27 (4)^0 (0)^13 (2)Females 47 (7)^0 (0) 6 (1)n=1576subjects held their legs off of the examination table. Nosignificant main effects or interactions were found.Analyses of Demographic Variables Using Chi Square analyses categorical demographic variableswere analyzed by group and by sex. No differences were foundamong groups or between sexes with respect to marital status(married as compared to single, separated, divorced, or widowed),first language spoken (English as compared to German, Hungarian,Italian, East Indian, or Native Indian) or work status (workingor carrying out regular activities if housewife or if retired ascompared to not working). Compensation status was not found todiffer between sexes, but did differ among groups, X 2 (2)=19.20,< .0001.^Multiple comparisons revealed that there were fewerpatients in the chronic congruent group compared to both theacute, and chronic incongruent pain groups who were receivingcompensation (i.e., payments from the Workmen's CompensationBoard or the Insurance Corporation of British Columbia) ordisability payments. The frequency of occurrence of thecategoiical variables are presented in Table 7.Group and sex differences on age and socioeconomic statuswere analyzed using a MANOVA. The means and standard deviationsof these variables are presented in Table 8. A main effect forgroup, but not for sex was found, F (4, 168)=3.38, p < .05. Nointeractions between the variables were present. Univariateanalyses revealed that the effect held only for socioeconomicstatus, F (2, 84)=4.19, p < .05. Tukey's multiple comparisonsrevealed that patients in the chronic congruent group had aTable 7Group Means on Categorical Demographic VariablesVariableMarital StatusAcuteGroupCongruent IncongruentMarried 9 10 13 10 9 10Other 6 5 2 5 6 5First LanguageEnglish 13 11 12 15 13 12Other 2 4 3 0 2 3Work StatusReg.^Status 7 5 3 2 8 3Not Working 8 10 12 13 7 12CompensationNo 6 6 14 14 9 11Yes 9 9 1 1 6 4n=15Table 8Group Means on Age and Socioeconomic StatusVariable^ GroupAcute^Congruent^IncongruentM^F M^F M^FAge^M^38.5 39.4^48.0 44.3^46.9 47.6^SD 12.3 15.3 14.1 16.4 15.1 15.8SES^M^36.6 31.8^45.4 38.7^36.9 30.4SD^9.8 12.6 14.4 14.3 12.2 11.6n=15Table 9Group Means on Self Reported Job SatisfactionGroupAcute^Congruent^IncongruentM (14) F (9)^M (13) F (11) M (12) F (7)M^8.3^5.9^8.1^6.7^7.4^6.4SD 1.8^2.6 1.8 2.9 2.3 4.07778significantly higher socioeconomic status than patients in eitherthe acute pain group or the chronic incongruent pain group. Thedifferences between the later two groups were not significant.Finally, an ANOVA was carried out to determine whether therewere differences among groups and between sexes in self reportedjob satisfaction. Patients who were housewives or were retireddid not make this rating, and were therefore not included in thisanalysis. The reduced number of subjects precluded the use ofjob satisfaction as a variable in the above MANOVA since usingthis variable in that analysis would have reduced the number ofsubjects in the analysis of age and socioeconomic status as well.A sex difference, but not a group difference, in reported jobsatisfaction was found, F (1, 65)=6.42, p < .05. The means (seeTable 9) reveal that females reported significantly less jobsatisfaction than males. It is of note, however, that the meanreported job satisfaction was above average for both sexes.Analyses of Patient Pain Related Variables Using Chi Square statistics the patient groups were comparedon their reported regular use of medication, and on theirreported use of medication taken on the day of the study. Nodifferences among groups or between sexes were found with respectto whether they had taken muscle relaxants or opiate analgesicson the day of the study. A sex, but not a group difference wasfound in the use of anti-inflammatories on the day of the study.Approximately 20% (n=9) of females compared to 2% (n=1) of malestook this type of drug on the day of the study.79With regard to regular medication use, there were no sex orgroup differences found in the reported regular use of musclerelaxants. A group difference, but not a sex difference, wasfound in the reported regular use of opiate analgesics, X 2(2)=7.11, R < .05. Multiple comparisons revealed that the acuteand chronic incongruent pain groups reported more frequent use ofopiate analgesics than chronic congruent patients. Approximately17% (n=8) of patients in the acute pain group, and 17% (n=8) ofpatients in the chronic incongruent pain group reported usingopiate analgesics on a regular basis compared to only 2% (n=1) ofpatients in the chronic congruent group making such a claim.Finally a sex difference, but not a group difference wasfound in the reported regular use of anti-inflammatories, X 2(2)=8.40, R < .005. Once again more females (33%, n=15) ascompared to males (6%, n=3) reported using anti-inflammatories ona regular basis.Continuous variables relating to the pain condition (i.e.,physical impairment and perceived disability) were entered into aMANOVA with group and sex as the independent variables. Themeans and standard deviations are presented in Table 10. Both asignificant main effect for group (F (4, 168)=10.18, R < .0001),and for sex (F (2, 83)=4.82, R < .01) emerged.Univariate F tests revealed significant group differenceswith regard to physical impairment (F (2,84)=15.86, R < .0001),and disability (F (2,84)=3.77, R < .05). Tukey's multiplecomparisons of the differences in physical impairment showed thatsubjects in the chronic incongruent pain group had greaterTable 10Group Means on Physical Impairment and DisabilityVariable^ GroupAcute^Congruent^IncongruentM^F M^F M^FImpair^M^2.5^2.6^7.2^8.7^12.9 13.1SD 4.0^3.0 7.5^5.8 9.5 10.3^Disability M 29.1 38.8^18.4 32.3^32.5 37.2SD 13.2 13.2 13.2 16.2 14.6 17.9n=158081physical impairment than patients in both the acute and chroniccongruent pain group. The differences between the latter twogroups were also significant with subjects in the chroniccongruent group showing more physical impairment than subjects inthe acute pain group. Multiple comparisons of differences amonggroups in perceived disability revealed that the chronicincongruent pain group and the acute pain group did not differfrom each other in perceived disability, but did, however, differfrom the chronic congruent pain group, with the former two groupsreporting greater disability than the latter group. Univariate Ftests also revealed a significant sex difference, with regard todisability (F (1,84)=9.10, p < .003), but not with regard tophysical impairment. An examination of the means in Table 10reveals that the females tended to perceive themselves as moredisabled than the males.Relations Among Incongruent Pain Measures and SeverityThe relations between physical impairment and disability andthe incongruent pain measures has been previously evaluated(Waddell et al., 1984; Mahon, 1991; Reesor & Craig, 1988).Moderate correlations between disability and the incongruentmeasures have been found consistently (Mahon, 1991; Reesor &Craig, 1988). The relations between physical impairment and themeasures of incongruent pain have been somewhat less consistent.Reesor and Craig (1988) found physical impairment to bemoderately correlated with the pain drawing and the inappropriatesymptom inventory, but not with the nonorganic physical signsassessment. Mahon (1991) found physical impairment to correlate82with the inappropriate symptom inventory and not with the paindrawing and nonorganic signs assessment. In the present study,disability and physical impairment were moderately correlatedwith all of the measures. The correlations are presented inTable 11.Choice of CovariatesIn the above analyses examining group differences ondemographic and pain related variables, it was found that thegroups differed on socioeconomic status, compensation status,reported regular use of pain killers, perceived disability andphysical impairment. These variables were all considered fortheir potential use as covariates. It was decided that two ofthe variables (reported regular use of pain killers, andperceived disability) were unsuitable as covariates since theywere subjective and had a large degree of measurement error. Athird variable (compensation status) was also unsuitable as acovariate since it was a categorical variable and the adjustmentof group means would render the results incomprehensible (e.g.,the means would be adjusted such that the patients were neitheron or off compensation status). A fourth variable (physicalimpairment) was also considered unsuitable since it was biasedtoward the acute pain patients obtaining lower scores on thismeasure than the chronic pain patients.The variable that was left to consider as a potentialcovariate was socioeconomic status which was objectively scoredand had low measurement error. In the analyses which follow, thecovariate was correlated with the dependent variables ofTable 11Correlations Among Measures of Incongruent Pain and SeverityVariable Nonorganic Inappropriate Pain DrawingSigns Symptoms ScorePhysical .32^** .41^* .25^*ImpairmentSubjective .27^* .28^* .28^*Disabilityn=90*p < .01** .p. < .0018384interest, but no significant relations between the dependentvariables and covariate were found. Consequently, no covarianceanalyses were performed since to do so would only reduce thedegrees of freedom and result in a loss of power.Analyses of Cognitive and Affective Measures A MANOVA with group and sex as independent variables wasused to examine group and sex differences on the cognitive andaffective variables (see Method section for description of eachvariable). Means and standard deviations of these variables canbe found in Table 12. While no main effects for sex, orinteractions emerged, a significant main effect for group wasfound, F (14,158)=1.76, p <.05. Univariate F tests examiningdifferences among the groups on the seven cognitive and affectivevariables revealed differences among groups on: 1) the tendencyto employ passive coping strategies, F (2,84)=3.59, p < .03; 2)the tendency to report catastrophizing cognitions (F (2,84)=4.91,< .01); and 3) the tendency to respond emotionally to the pain,F (2,84)=6.35, p < .003. Tukey's multiple comparisons revealedthat acute and chronic incongruent patients did not differ intheir scores, but did differ from chronic congruent patients.That is, they scored significantly higher than chronic congruentpatients on all three measures.Analyses of Verbal Report of PainSince some subjects were only able to lift one leg up off ofthe examination table instead of two it was first of interest toexamine whether these subjects differed in the reported intensityand unpleasantness of the pain they experienced in response toTable 12Group Means and Standard Deviations on the Cognitive andAffective Variables VariableCognitiveAcuteM^FGroupCongruentM^FIncongruentM^FM 2.22 2.42 2.48 2.47 2.37 2.34SD 1.04 0.91 1.08 0.87 1.08 1.09EvaluativeM 3.03 3.37 2.87 3.47 2.87 2.83SD 1.24 1.17 1.09 1.03 1.20 1.05PassiveM 2.08 2.95 1.57 1.93 2.41 2.33SD 1.07 1.30 1.18 0.99 1.14 1.31CatastrophicM 1.12 1.91 0.63 1.08 1.46 2.02SD 0.82 1.48 0.74 0.95 1.32 1.31BehaviouralM 2.66 3.00 2.51 3.07 2.90 2.87SD 1.0 1.16 1.55 1.16 1.26 1.57WorryM 2.99 3.26 2.61 2.89 2.91 4.02SD 0.91 1.47 1.68 1.13 1.58 1.45EmotionalityM 2.43 3.12 1.72 2.03 2.34 3.14SD 1.01 1.05 0.92 1.05 1.28 1.27n=158586the physiotherapy examination. No differences were observed, andwe were therefore reasonably confident that this slight deviationfrom the experimental procedure would not confound the results.A repeated measures MANOVA was carried out to determinewhether there were group or sex differences in the intensity andunpleasantness of the pain experienced either on a daily basis orin response to the painful range of motion exercises. The meansand standard deviations of the variables are presented in Table13. The MANOVA revealed that there was an overall effect ofgroup (F (4,168)=2.57, R < .05), and time (daily versusmovement), (F (2,83)=12.92, E < .001). No main effects involvingsex were found, and in addition, no interactions among thevariables were present.Univariate F tests revealed that there was a significantdifference among the groups in the reported unpleasantness ofpain, (F (2,84)=5.12, R < .008), but not in the reportedintensity of pain. Tukey's post hoc tests revealed that thechronic incongruent patients reported their pain to besignificantly more unpleasant than the chronic congruent patientsbut not the acute pain patients. No differences between theacute and chronic congruent pain patients were observed.Univariate F tests also revealed that the pain reported on adaily basis was significantly greater than the pain that wasreported in response to the movement both in terms of intensity(F (1,84)=14.99, R < .0001) and unpleasantness (F (1,84)=25.42,R< .0001). An examination of the means revealed that the painTable 13Group Means and Standard Deviations on Pain Intensity andUnpleasantnessVariable GroupAcute Congruent IncongruentM F M F M FDaily PainInten. M 21.54 27.21 22.82 23.35 28.20 32.91SD 11.78 17.93 10.29 12.53 15.33 15.73Unpl. M 14.65 18.69 10.75 14.77 18.65 19.30SD 10.50 12.39 6.77 11.21 13.31 10.88MovementInten. M 13.67 20.69 14.17 20.07 17.23 30.78SD 15.67 17.14 12.08 15.68 16.80 16.33Unpl. M 8.49 5.81 6.61 7.03 10.32 19.29SD 11.83 3.70 5.95 6.18 14.51 14.84n=158788experienced on a daily basis was reported to be significantlymore intense and unpleasant than the pain experienced inresponse to the leg movement.Analyses of Facial ActivityBecause of the large number of AUs available to study (e.g.,40) it was decided to only include in the analyses those AUswhich either: 1) were found to be related to pain in previousstudies of pain (Craig, Hyde & Patrick, 1991); or 2) wereobserved to occur more than 5% (14/270) of the time across allpain segments (i.e., genuine, masked, and exaggerated) and acrossall subjects. The AUs retained for the analyses were AUs: 1, 2,4, 6 and 7 combination, 9 and 10 combination, 12, 17, 20, 24,25,26 and 27 combination, 43, and 45 (see Appendix L for a briefdescription of the AUs or AU combinations.To begin, a repeated measures MANOVA was carried out withthe frequency of the facial variables as the dependent variables,group and sex as the independent variables, and instructional set(i.e, baseline, masked, genuine, and exaggerated) as the repeatedmeasure. There were no main effects or interactions involvingeither group or sex. There was, however, an overall main effectfor instructional set, F(36, 729)=4.19, p < .0001. Univariate F-tests (see Table 14) revealed that there were significantdifferences among the instructional sets for all variables exceptAUs 2 (outer brow raise), 24 (lip press), and 45 (blink). Themeans of the variables are presented in Table 15, and a summaryof Tukey's post hoc tests are presented in Table 16.89Table 14Univariate F-tests on Frequency of AUs Across Instructional Sets.Variable01-Inner Brow Raise 2.94 .0304-Brow Lowerer 15.77 .000106/07-Cheek Raise/Lids Tight 21.88 .000109/10-Nose Wrinkle/Upper Lip Raise 5.84 .00112-Lip Corner Pull 5.06 .00217-Chin Raise 3.12 .0320-Lip Stretch 7.24 .000143-Eyes Closed 15.63 .000125/26/27-Mouth Opening 7.44 .0001Table 15Mean Frequency of Occurrence of the Variables AcrossInstructional Sets.Baseline Masked Genuine Exag.Variable01-Inner Brow Raise .06 .08 .10 .2004-Brow Lowerer .02 .09 .14 .3306/07-Orbit Tightening .03 .19 .32 .6409/10-Levator Contract .00 .02 .12 .1712-Lip Corner Pull .07 .04 .13 .2217-Chin Raise .02 .10 .11 .1720-Lip Stretch .03 .13 .06 .2625/26/27-Mouth Opening .20 .24 .56 .6243-Eyes Closed .01 .07 .08 .1990Table 16Results of Tukey's Post Hoc Tests on Differences in the Frequencyof AUs Across Instructional Sets.Instructional Set^Differences In Frequency of AUsBaseline versus:MaskedGenuineExaggeratedMasked versus: No differences.AUs 4, 6+7, 9+10, 25+26+27 occurred morefrequently during the genuine segment.AUs 1, 4, 6+7, 9+10, 12, 17, 25+26+27, 43occurred more frequently during theexaggerated segment.Genuine^AU combination 25, 26, 27 occurred morefrequently during the genuine segment.Exaggerated^AUs 4, 6+7, 9+10, 12, 25+26+27, 43occurred more frequently during theexaggerated segment.Genuine versus: Exaggerated^AUs 4, 6+7, 20, and 43 occurred morefrequently in the exaggerated painsegment.91A second repeated measures MANOVA was carried out to examineif there were group, sex, or instructional set differences in themean intensity of the AUs. There were no main effects orinteractions involving either group or sex. Once again, anoverall main effect for instructional set, F(36, 729)=4.67, p <.0001, was found. Univariate F-tests (see Table 17) revealedthat there were significant differences among the instructionalsets for all AUs except 45 (blinking). The means of thevariables are presented in Table 18, and a summary of Tukey'spost hoc analyses are presented in Table 19.The frequency scores of those variables which were found tooccur more frequently during the genuine pain segment as comparedto the baseline segment (i.e., AU 4, 6+7, 9+10, and 25+26+27)were entered into a principal components analysis (PCA) todetermine how the separate pain related variables interrelatedempirically. This resulted in the emergence of one factor whichaccounted for approximately 44 percent of the variance in facialactivity. The factor loadings are presented in Table 20.Prediction of Pain GroupThose variables which were found in the above analyses todiffer with respect to group were entered into a discriminantfunction analysis to determine whichdiscriminated among the pain groups.included all of those variables fromwere found to be reliably related tostatus, socioeconomic status, use ofdisability, physical impairment, usevariables optimallyThe predictor variablesthe above analyses whichgroup status: compensationopiate analgesics, perceivedof passive copingTable 17Results of Univariate F-tests Examining Mean Intensity of AUsAcross Instructional Sets.F PiVariable01-Inner Brow Raise 7.00 .000102-Outer Brow Raise 4.38 .00504-Brow Lowerer 18.30 .000106/07-Cheek Raise/Lids Tight 34.10 .000109/10-Nose Wrinkle/Upper Lip Raise 8.49 .000112-Lip Corner Pull 5.65 .00117-Chin Raise 4.01 .00820-Lip Stretch 12.69 .000124-Lip Press 3.61 .01443-Eyes Closed 12.99 .000125/26/27-Mouth Opening 8.15 .000192Table 18Mean Intensity of AUs Across Instructional Sets.Exag.Variable Baseline Masked Genuine01-Inner Brow Raise .09 .09 .19 .4902-Outer Brow Raise .07 .04 .07 .3004-Brow Lowerer .03 .13 .32 .8706/07-Orbit Tightening .10 .36 .53 1.509/10-Levator Contract .00 .04 .18 .3912-Lip Corner Pull .17 .09 .26 .5417-Chin Raise .06 .13 .20 .3420-Lip Stretch .03 .19 .09 .5924-Lip Press .01 .06 .18 .1625/26/27-Mouth Opening .29 .40 .71 .8243-Eyes Closed .01 .12 .13 .4393Table 19Results of Tukey's Post Hoc Tests Examining Differences in theMean Intensity of AUs Across Instructional Sets.Instructional Set^Differences In Frequency of AUsBaseline versus:Masked^No differences.Genuine AUs 6+7, 24, and 25+26+27 were greaterin intensity during the genuine segment.Exaggerated^The mean intensity of all Aus was greaterduring the exaggerated segment.Masked versus: Genuine^AU combination 25, 26, 27 was moreintense during the genuine segment. Noother differences were found.Exaggerated^All of the AUs were of greater intensityduring the exaggerated segment.Genuine versus: Exaggerated^AUs 1, 2, 4, 6+7, 9+10, 20, and 43 wereof greater intensity during theexaggerated segment.Table 20Factor Loadings of Pain Related AUs.Variable^ Loading^Total Variance04-Brow Lowerer^ .75 44%06/07-Orbit Tightening^.7109/10-Levator Contraction .6825/26/27-Mouth Opening .5094strategies, presence of catastrophizing cognitions, tendency torespond emotionally to the pain, unpleasantness of the pain on adaily basis, and in response to the physiotherapy movement.Two significant discriminant functions emerged: 1) X 2(20)=77.09 2 < .00001; 2) X 2 (9)=22.09, p < .009). The twodiscriminant functions accounted for 49% and 23%, respectively,of the between group variability. By examining the group meanson the functions (see Table 21) it is clear that the firstdiscriminant function maximally separated acute pain patientsfrom the chronic pain groups, while the second discriminantfunction maximally separated chronic incongruent pain patientsfrom chronic congruent pain patients, with acute pain patients inbetween.The loading matrix of correlations between predictors anddiscriminant functions (see Table 22) suggests the best predictorfor distinguishing between acute and chronic pain groups isphysical impairment (in line with Tabachinick and Fidell (1989)only loadings greater than .50 were interpreted). This is notsurprising since the physical impairment of the patient is inpart determined by the duration of the pain complaint, and thenumber of surgeries both of which are greater for chronic ascompared to acute pain patients.Further examination of Table 22 suggests that the bestpredictors for distinguishing between chronic incongruent andchronic congruent with acute pain patients in between are: 1)emotional responses to pain; 2) presence of catastrophizingcognitions; 3) reported unpleasantness of the leg movement;Table 21Group Means on the Discriminant FunctionsGroup^ Function 1^Function 2Acute -1.35 0.02Chronic Congruent^0.65^ -0.68Chronic Incongruent 0.70 0.66Table 22Results of the Discriminant Function Analysis Predicting GroupStatus.Predictor Variables 1 2 UnivariateF^(2,87)Impairment 0.56 0.51 16.36 ***Emotionality -0.20 0.58 6.05 **Catastrophizing -0.09 0.57 4.72 **Movement Unpleasantness 0.18 0.55 5.40 **Socioeconomic Status 0.13 -0.50 4.08 *Compensation Status -0.46 0.49 11.80 ***Disability -0.12 0.47 3.47 *Regular Pain Killers -0.14 0.46 3.73 *Day Unpleasantness -0.03 0.43 2.46 *Passive Coping -0.18 0.40 3.52 *Canonical R 0.70 0.48Eigenvalue 0.95 0.31*** p <^.0001;^**^p < .01;^* p < .05Table 23Classification Results Based on the Two Discriminant Functions. Group^ 1^2^3Acute 25^5 0(83.3%)^(16.7%)^(0.0%)Chronic Congruent^4 20 6(10.0%)^(66.7%)^(23.3%)Chronic Incongruent^3 8 19(13.3%)^(23.3%)^(63.3%)95964) physical impairment; and 5) socioeconomic status.Consideration of both discriminant functions resulted in correctclassification of the groups 71% of the time. More specificclassification results are presented in Table 23.Prediction of Membership in Chronic Pain Groups The distinction between acute and chronic pain groups hastraditionally been made on the basis of the duration of the paincomplaint, and is relatively straightforward. The distinctionbetween chronic congruent and incongruent pain patients, however,is much more complicated since health care professionals mustoften rely on multiple sources of information in making such adistinction. Since this distinction is more complex it wasdecided to carry out another discriminant function analysisdisregarding information on the acute pain patients and usingonly information on the chronic pain groups. This analysis alsoallowed for a more direct comparison with other research in thefield (e.g., Reesor & Craig, 1988). Once again those variableswhich were found in the above analyses to differ with respect togroup were entered into a discriminant function analysis todetermine which variables optimally discriminated among thechronic congruent and incongruent pain groups.One significant discriminant function emerged (X 2 (10)=20.922 < .05), and accounted for 32% of the between group variability.The mean on the function for chronic congruent patients was .68,whereas the mean on the function for chronic incongruent patientswas -.68. The loading matrix of correlations between thepredictors and the discriminant function (see Table 24) suggestedTable 24Results of Discriminant Function Analysis Predicting Congruentand Incongruent Pain Groups.Function0.580.540.510.510.50-0.460.440.430.420.390.570.481 UnivariateF^(2,87)^9.35^**8.27^**7.23^**7.25^**6.93^**5.93^*5.55^*5.21^*5.05^*4.33^*Predictor VariablesCatastrophizingEmotionalityMovement UnpleasantnessCompensation StatusRegular Pain KillersSocioeconomic StatusPhysical ImpairmentDisabilityDay UnpleasantnessPassive CopingCanonical REigenvalue** 2 < .01* 2 < .05Table 25Classification Results of Congruent and Incongruent Patients. 971 224 6(80.0%) (20.0%)11 19(36.7%) (63.3%)GroupChronic CongruentChronic Incongruent98that the best predictors of group status were: 1) variablesrelated to the negative interpretation of pain (i.e.,emotionality, catastrophizing, movement unpleasantness); 2)compensation status; and 3) reported regular use of pain killers.That is, chronic incongruent patients could be discriminated fromchronic congruent patients in terms of their increased likelihoodto negatively interpret their pain, be on compensation, and usepain killers on a regular basis. Consideration of thediscriminant function resulted in correct classification of thegroups 72% of the time. More specific classification results arepresented in Table 25.Predictors of Verbal and Nonverbal Measures of Pain.Because in general the verbal and nonverbal measures of painwere of limited use in discriminating among groups, it was ofinterest to examine which variables collected in the study werein fact related to the verbal and nonverbal measures of pain.The criterion variables in three separate sets of stepwiseregression analyses were: 1) the intensity of pain reported bythe subjects in response to the painful movements; 2) theunpleasantness of pain reported by the subjects also in responseto the movement; and 3) a composite genuine painful facialactivity score created by summing the frequency of action units4, 6 and 7, 9 and 10, and 25,26 and 27 that resulted whensubjects were asked to genuinely express their pain.For each criterion seven different stepwise multipleregression analyses were carried out. One analysis was carriedout for each of six sets of predictors. In addition, one final99analysis was carried out which used as predictors those variableswhich from the previous analyses were found to be significantlyrelated to the criterion. This procedure, although somewhatcomplicated, served to reduce the ratio between number ofsubjects and predictors for each regression analysis.^The sixgroups of predictor variables were as follows:1) group assignment measures - the duration of paincomplaint, the pain drawing, the inappropriate symptomsinventory, and the nonorganic signs score.2) demographic variables - marital status, English as firstor second language, work status, compensation status, age,socioeconomic status.3) pain related variables - medication taken on a regularbasis or the day of testing, physical impairment, anddisability.4) cognitive and affective variables - use of cognitive,evaluative, passive, and behavioural coping strategies,presence of dysfunctional cognitions, worry, andemotionality.5) verbal report of pain - the intensity and unpleasantnessof daily pain and movement pain.6) nonverbal expression of pain - the sum of the painrelated activity that occurred during baseline, masked,genuine, and exaggerated pain segments.Predictors of Movement Pain Intensity. Of the groupassignment measures, the pain drawing (B=.38, 2 < .001) and thenonorganic signs assessment (A=.31, 2 < .002) were found to be100significant predictors of the reported intensity of the painprovoked by the movement. Together they accounted for 24% of thevariance in the reported pain intensity. Of the demographicvariables first language spoken, (B=.26, R < .03), andsocioeconomic status (B=-.21, R < .04) both significantlypredicted pain intensity. Together the variables accounted for11% of the variance in reported pain intensity. Regular use ofopiate analgesics (B=.27, P < .009) also related to the intensityof the painful movement accounting for 17% of the variance inreported pain intensity.From the cognitive and affective variables, tendency tocatastrophize (B=.29, p < .009) and to use behavioural copingstrategies (E=.26, p < .009) predicted pain intensity andaccounted for 14% of the variance. The intensity of the movementwas also significantly related to the reported unpleasantness ofthe movement (B=.77, R < .00001) and the reported intensity ofpain experienced on a daily basis (B=.22, R < .002). Thevariables together accounted for 64% of the variance in painintensity. Of the nonverbal expressive measures, only the sum ofpain related variables during the genuine pain segment (B=.42,< .0001) significantly related to the reported pain intensityaccounting for approximately 18% of the variance in painintensity scores.Variables which in the above analyses were significantlyrelated to reported pain intensity were entered into a furtherstepwise regression analysis. The results of this analysis arereported in Table 26. The best predictors of pain intensity wereTable 26Significant Predictors of Movement Intensity.Step^Variable^ Beta^p<^Total R 21^Unpleasantness of^.77^.00001^.66Movement2^Intensity of Daily^.22^.03Pain3^Behavioural Coping^.15^.01Strategies4^Pain Drawing^.17^.007Measure5^Genuine Pain^.13^.04Expression1.01102other verbal report measures of pain, the use of behaviouralstrategies, the pain drawing, and the amount of genuine facialactivity.Predictors of Movement Unpleasantness. Of the groupassignment measures, the nonorganic signs assessment (B=.36, 2 <.002) and the pain drawing (k =.20, 2 < .05) accounted for 17% ofthe variance in the reported unpleasantness of the movement.Compensation status (B=.27, 2 < .009) proved to be a significantpredictor of pain unpleasantness accounting for approximately 8%of the variance. The use of anti-inflammatories on a regularbasis (B=.25, 2 < .02) and physical impairment (B=.23, R < .03)also together accounted for 11% of the variance in the reportedunpleasantness of the movement.Of the cognitive and affective variables, worry (11=.21, 2 <.05) was the only variable which accounted for a significantproportion of the variance (R 2=.05) in pain unpleasantness. Theintensity of the movement (B=.77, 2 < .00001) was alsosignificantly related to pain unpleasantness, accounting forapproximately 59% of the variance. Finally, among the painexpression measures facial activity during the genuine painsegment (B=.41, R < .0001) accounted for a significant proportionof the variance in reported pain unpleasantness (R2 =.16).Variables which related to pain unpleasantness in the aboveanalyses were entered into a further regression analysis aspredictor variables. Only reported pain intensity entered intothe equation to significantly predict pain unpleasantness (seeTable 27).Table 27Significant Predictors of Movement Unpleasantness.Step^Variable^ B^p<^Total R 21^Movement Intensity^.77^.00001^.59103104Predictors of Genuine Painful Facial Activity. Of the groupassignment measures only the pain drawing (B =.32, 2 < .002)significantly predicted (R2 =.10) the painful facial activityobserved during the genuine pain segment. Of the demographicbackground variables, marital status (B=.23, p < .03) was foundto be significantly related to the genuine expression of painaccounting for approximately 5% of the variance. The expressionof genuine pain was also positively associated with the reportedsensory intensity of the movement (B = .42, 2 < .00001) whichaccounted for 18% of the variance in the facial activity score.Finally, the genuine expression of pain was significantly relatedto the exaggerated expression of pain (B = .43, 2 < .00001), andthe amount of painful facial activity observed during thebaseline segment (B=.30, 2 < .002). Together these variablesaccounted for 28% of the variance in the genuine pain expression.Once again, those variables which were significantly relatedto the genuine expression of pain score in the above analyseswere entered together as predictors into a stepwise multipleregression analysis. The results of this analysis (see Table 28)suggest that the degree to which subjects expressed pain in thegenuine pain segment was largely a function of how much they werewilling to exaggerate their pain, and how much facial activitythey showed during baseline. In addition, however, the degree ofgenuine painful facial activity was also related to how intensesubjects found the painful leg movement.Table 28Significant Predictors of the Pain Expression During the GenuinePain Segment. 105Step123VariablePain Expression Duringthe Exaggerated SegmentMovement IntensityPain Expression Duringthe Baseline SegmentBeta^Total R 2.43^.0003^.34.32^.004.30^.005DISCUSSIONOverview of DiscussionThe results of the present study suggested that acute andchronic incongruent pain patients had a similar psychologicalreaction to pain that was greater than the psychological reactionto pain of chronic congruent patients. In addition, the resultssuggested that acute patients and chronic incongruent patientsreported their pain to be similar in unpleasantness, althoughonly the chronic incongruent patients had significantly higherscores on this measure than the chronic congruent patients.Acute patients and chronic incongruent patients were also moresimilar to each other than they were to the chronic congruentpain groups on several demographic variables (socioeconomicstatus, compensation status) and several pain related variables(regular use of opiate analgesics, and disability). One finaldifference among groups was observed on a measure of physicalimpairment in which it was found that the chronic incongruentgroup had greater physical impairment than the chronic congruentand acute pain groups. Differences between the latter two groupswere also significant with the chronic congruent patients showingmore physical impairment than the acute pain group.Of all the variables noted above the most important variablefor optimally discriminating among acute and chronic pain groupswas physical impairment. This, however, may have been anartifact since the physical impairment measure was biased towardchronic pain patients obtaining higher scores. The mostimportant variables for discriminating among chronic congruent106107and incongruent pain groups were related to the patient'snegative interpretation of pain, compensation status, andreported regular use of pain killers.No differences among groups were observed in their nonverbalexpressions of pain. This, however, may have been an artifact ofthe procedures used to study facial expressions of pain. Thatis, having subjects not only genuinely express pain, but alsomask and exaggerate their pain expressions may have served toattenuate the facial actions observed, and as a resultdifferences among groups may have been obscured.In general, the facial actions that patients displayed whenthey were asked to genuinely express pain in response to the legmovements corresponded to facial actions that in previous studieshave been found to be associated with pain, and included browlowering, orbit tightening, levator contraction, and mouthopening. Also consistent with previous research were findingssuggesting that, although subjects were remarkably successful inboth masking and exaggerating pain, there were still some cues todeception apparent on the face.Clarification of Group Status It should be recalled that patients were assigned to theacute and chronic pain groups on the basis of whether they hadexperienced LBP for a period less than or greater than threemonths. This distinction was effective in separating the groups,and resulted in the formation of an acute pain group who had onaverage experienced their pain for a period less than one month,and the formation of the two chronic pain groups who had on108average experienced their pain for a period of six and a halfyears.It should also be recalled that chronic pain patients werefurther subdivided into congruent and incongruent pain groups.Patients were assigned to the incongruent pain group if they had:1) a score of two or greater on the nonorganic signs assessment;or 2) a score - of three or greater on the inappropriate symptominventory, or 3) a score of five or greater on the pain drawing.When patients did not meet these criteria they were assigned tothe chronic congruent pain group.Using these criteria it was found that only one of thechronic incongruent patients actually met all three of thecriteria, about half of the incongruent patients met two of thecriteria, and about half only met one of the criteria. Thesefindings are similar to those of Reesor and Craig (1988) andMahon (1991) who also used this classification system.Also of interest in the present study is the fact that thepain drawing was of limited use in classifying patients asincongruent. When patients met two of the criteria, they weremore likely to have met the criteria for the nonorganic signsassessment, and the inappropriate symptom inventory, and whenpatients had met only of the criteria, it was more likely to havebeen the criteria for these measures as well, and not the paindrawing. Reesor and Craig (1988), and Mahon (1991) did not findthis. That is, they found all three measures to be equallyeffective in identifying the incongruent pain patients. Thedifferences between the studies may be a result of the different109patient populations. In the studies conducted by Reesor andCraig (1988) and Mahon (1991) the patients were drawn from acomprehensive back pain assessment center providing tertiary carewhich for many patients was the last resort for help. Incontrast, the patients in the present study were drawn from aprimary care physiotherapy clinic, which was likely to be one ofthe patients' first resorts for help.Despite the fact that the pain drawing was not usedfrequently to classify patients as incongruent, patients in theincongruent pain group still scored higher on average thanpatients in the chronic congruent group on this measure as wellas on the other measures. This may suggest that to properlyidentify patients as chronic incongruent on the basis of the paindrawing (at least within a physiotherapy clinic setting) thecriterion score should be lowered perhaps to the mean of 3.While the acute pain patients were not further subdividedinto groups on the basis of the incongruent pain measures, it wasstill of interest to determine how they scored on these measuresrelative to the chronic pain groups. When this comparison wasmade it was found that the acute pain patients were more similarto the chronic congruent patients than to the chronic incongruentpatients on the pain drawing, and on the inappropriate symptominventory. On the nonorganic signs assessment, however, thescores of the acute pain patients fell between the scores of thechronic congruent and incongruent subjects.Overall with respect to the classification criteria, itwould appear that the acute pain patients were more similar to110the chronic congruent pain patients than to the chronicincongruent patients. However, when we examined how many of thepatients in the acute pain group actually met the criteria foridentifying incongruent pain we found that six males and eightfemales met the criteria for incongruent pain presentation.Because of the small number of subjects involved it was notpossible to determine whether there were significant differencesbetween the acute pain groups on any of the dependent variablescollected in the study.Future research perhaps could be directed toward anexamination of difference among groups who show congruent andincongruent signs in the acute pain stage. It may be thatpatients who show incongruent signs in the acute pain stage ascompared to patients who show congruent signs in the acute painstage are more likely to have their pain develop into a chronicpain condition. In general, it would seem important to determinewhether patients who show incongruent signs in the acute painstage are the same patients who show incongruent signs in thechronic pain stage. If the signs persist it may suggest that onecan identify patients with incongruent pain much earlier in thepain cycle than was previously thought possible, and this may inturn have implications for treatment. That is, since patientswith incongruent pain have poorer coping strategies and moredysfunctional cognitions than congruent pain patients (Reesor &Craig, 1991) the treatment package for acute pain patients withincongruent signs may be more effective if, in addition to the111routine focus on physical treatment modalities, more emphasiswere also placed on psychological treatment strategies.In general, few sex differences on the incongruent painmeasures were found. Unlike the studies conducted by Reesor andCraig (1988) and Mahon (1991), females as compared to males werenot found to be more likely to have met two or more theincongruent criteria. In addition, the mean scores on the paindrawing, and the inappropriate symptom inventory were not foundto differ between males and females.The exception to finding no sex differences was on thenonorganic signs assessment measure where it was found thatfemales in the chronic incongruent pain group obtained higherscores than males in that same group. In addition, it was foundthat females in the chronic incongruent group had higher scoresthan both acute and chronic congruent patients, whereas males inthe chronic incongruent only had higher scores than the chroniccongruent pain group and not the acute pain group.Upon further examination of the nonorganic signs inventoryit was found that the females obtained higher scores on thisinstrument because they more frequently exaggerated theirexpressions of pain in response to examination, and reported painin response to axial loading of the spine. The scoring ofexaggerated expressions of pain in this instrument is likely verysensitive to observer biases, and as a result the finding thatwomen score higher on this instrument needs to be treatedcautiously.112Group Differences Demographic Variables. In the present study the pain groupswere found to differ with respect to two demographic variables,namely compensation status and socioeconomic status. In bothcases, chronic incongruent patients and acute pain patients werefound to be more likely on compensation and to be of a lowersocioeconomic status compared to chronic congruent patients.Research comparing acute and chronic pain patients isparticularly sparse (Ackerman & Stevens, 1989), and as a resultit is difficult to determine whether within this sample the acutepain patients are unusual or usual with respect to theirsimilarity to chronic incongruent patients and dissimilarity tochronic congruent patients.Research concerning chronic congruent and incongruentpatients is more abundant. At the same time, however, comparisonof the present findings to other findings is made difficult bythe fact that the chronic patients in other studies tend to besampled from tertiary health care settings rather than a moreheterogeneous community sample. These differences in samplingmay account for the fact that differences between chroniccongruent and incongruent patients on socioeconomic status andcompensation status were found in this study, but not in studiesconducted by Reesor and Craig (1988), Mahon (1991) and Waddelland colleagues (1980).There appear to be two alternative explanations for findingthat patients who have incongruent pain are more likely to comefrom a lower socioeconomic stratum, and be on compensation as113compared to patients with chronic congruent pain. Oneexplanation may be that patients begin to display incongruentsigns of pain when they are not coping well with their pain, andare under stress. The incongruent signs under this model may beviewed as either a result of the patient's emotionalvulnerability or hypersensitivity to pain and or the patient'sway of obtaining the attention of his or her health care providerin order that his or her pain condition be taken seriously. Inthe present context, negative life circumstances, such as beingon compensation or coming from a lower socioeconomic stratum, mayresult in increased stress and difficulties with coping, and as aconsequence the presence of multiple incongruent signs andsymptoms may become more likely. Recent research does supportthe notion that being on compensation is a major source of stressthat adds to the deterioration of the pain condition (Guest &Drummond, 1992).An alternative explanation for these findings is thatpatients, perhaps as a result of different learning histories,are predisposed to display their pain in certain ways. In thepresent context, this would suggest that patients who come from alower socioeconomic stratum and are likely to seek compensationmay have different learning histories from patients with highersocioeconomic status, and as a result display their pain in a waythat is incongruent with underlying anatomy and physiology.Unfortunately, the present results do not shed light onwhether congruent and incongruent symptoms occur as a result ofsome predisposition to express pain in this way, or as a result114of stress. It rests with future research to develop alongitudinal design to address this question.Pain Related Characteristics. The patient groups were alsofound to differ with respect to several pain related variables,namely physical impairment, perceived disability and reportedregular use of opiate analgesics. With regard to physicalimpairment, it was found that the acute pain patients (2.6%)showed the lowest degree of physical impairment, followed bychronic congruent patients (8%), and chronic incongruent patients(13%). All differences were significant.The mean score on the physical impairment index obtained bythe acute pain group is particularly notable, since itcorresponds to the scores that patients with no physicalpathology typically obtain. This suggests that the validity ofthe physical impairment inventory with acute pain patients isquestionable. The mean score of the chronic congruent patientgroup is also below what is expected in a LBP patient population,whereas the mean score of the chronic incongruent group justfalls within the expected range which Waddell (1987) reports tobe 10-20%. The scores on the physical impairment index in thisstudy are substantially lower than the scores obtained in a studyby Reesor and Craig (1988). Their chronic congruent subjects hada mean score of 13, while their chronic incongruent subjects hada mean score of 20 on physical impairment. Once again, thediscrepancy in scores may well be a result of sampling.It was not particularly striking to find that acute painpatients obtained significantly lower scores the chronic painuspatients on this measure. This follows from the fact that scoreson the physical impairment inventory are in part a function ofchronicity of pain complaint and number of previous surgeriesboth of which are higher for chronic pain patients than acutepain patients.What was striking in this study and in other studies (Reesor& Craig, 1988) was that chronic incongruent pain patients had ahigher level of physical impairment than chronic congruent painpatients. One explanation for this finding is that measures ofincongruent pain are actually measures of physical impairment.Although this may be the case, it seems implausible for a numberof reasons. First, patients with acute pain had the lowestlevels of physical impairment, yet their scores on theincongruent pain measures were higher than the congruent painpatients who had higher levels of physical impairment, andsignificantly so on at least one measure (i.e., nonorganicphysical signs measure). Second, it is unlikely that adifference in physical impairment of only 5% could produce suchsignificantly different incongruent pain scores. Third, it seemshighly improbable that by measuring nonorganic signs (e.g., thereport of pain in response to axial loading), inappropriatesymptoms (e.g., the presence of pain in the tip of the tailbone),or pain drawings (e.g., the nonanatomical presentation of pain inhighly exaggerated pain drawings) one is actually measuringphysical impairment.What then can account for the finding that patients who showmultiple incongruent signs also have higher levels of physical116impairment than patients who show few incongruent signs? Oneexplanation is that any variable (whether it be physicalimpairment, recent onset of the pain condition, poor copingstrategies, or negative life circumstances) which increases thepatient's uncertainty about his or her pain condition (e.g., interms of diagnosis, or effectiveness of treatment) anddifficulties in coping will result in an increase in the numberof incongruent signs and symptoms. The increase in incongruentsigns and symptoms can then be viewed as a result of either thepatient's excessive focus of attention on somatic functioning,and or the patient's way of getting a health care professional totake their pain condition seriously. Uncertainty and concern forone's health will not only likely result in an increased numberof incongruent signs, but may also result in decreased physicalactivity. The decreased physical activity over time may in turnresult in higher levels of physical impairment.An alternative explanation for the finding is that patientsare perhaps predisposed to respond to their pain in a manner thatis either congruent or incongruent with underlying anatomy andphysiology. If they respond incongruently this also suggeststhat they will have a greater psychological reaction to theirpain. The greater psychological reaction to pain could in turnresult in less activity, and as a result increased scores onmeasures of physical impairment. Both of the above explanationsappear plausible, and as a result future research which followsthe development of LBP longitudinally is needed to determinewhether 1) increased physical impairment causes increased stress117and as a result chronic incongruent pain presentation, decreasedphysical activity and further physical impairment; or 2) chronicincongruent pain which involves an increased psychologicalreaction to pain and decreased activity result in increasedlevels of physical impairment.Further differences among groups were also found on reporteddisability and use of opiate analgesics. Here it was found thatacute pain patients and chronic incongruent patients had agreater likelihood of using opiate analgesics on a regular basisand showed more functional disability than chronic congruent painpatients. More specifically, with respect to medication use, 17%of the acute pain patients, and 17% of the chronic incongruentpatients were using opiate analgesics on a regular basis comparedto only 2% of the chronic congruent patients. With respect todisability, while there were group differences, all of thepatient groups had scores between 20-40% which according toFairbank and colleagues (1980) suggests that patients weremoderately disabled (e.g, likely having greatest difficultiessitting, lifting, and standing, moderate difficulties withtravelling, socializing and working, and minimal difficultieswith personal care, sexual activity, and sleeping). The level ofdisability in the patients in this sample was comparable to thelevel of disability found in LBP patients in other studies(Mahon, 1991; Reesor & Craig, 1988).What is of particular interest here is not that acute painpatients show this level of disability and medication use, butthat chronic incongruent patient's do. That is, with acute pain118health care professionals often suggest that patients reducetheir activity, and use opiate analgesics to reduce pain. Withchronic pain, however, it is widely recognized that such actionsare often contraindicated in the reduction of pain (Bonica,1980), and thus what is surprising about these results is thatchronic incongruent patients are showing the same amount ofinactivity and medication use as the acute pain patient. Whythis is so can only be speculated upon. It can only be assumedthat the health care professionals who are treating thesepatients, not knowing what else to do for their patients,continue to prescribe medication, and suggest decreased physicalactivity.Cognitive and Affective Measures. Interesting groupdifferences, but not sex differences, on the cognitive andaffective variables also emerged. These differences suggestedthat acute and chronic incongruent pain patients responded moreemotionally to their pain, were more likely to use passive copingstrategies, and more frequently reported dysfunctional,catastrophizing cognitions than chronic congruent pain patients.Findings showing that congruent and incongruent painpatients differed were expected since other researchers havefound cognitive and affective variables to be important indiscriminating among these pain groups. For instance, Reesor andCraig (1988) found strong correlations among ineffective copingstrategies, maladaptive and dysfunctional cognitions, and highlevels of anxiety in patients with incongruent pain as comparedto patients with congruent pain (Reesor & Craig, 1988). Mahon119(1991) also investigated this relation, and while she did notfind significant differences between congruent and incongruentpatients, she did find a trend that suggested that patientsdiffered on cognitive appraisal and affective distress. Otherinvestigators have found dysfunctional personality traitsidentified by subscales on the MMPI (Doxey et al., 1988; Lehmannet al., 1983; Waddell et al., 1980), and depression (Main &Waddell, 1982) to be more strongly related to multiple nonorganicsigns than to few nonorganic signs.Findings showing that acute pain patients had similartendencies to chronic incongruent patients in their emotionalreactions to pain, use of passive coping strategies and tendencyto employ dysfunctional catastrophizing cognitions were notexpected. Rather, it was expected that patients whose pain hadpersisted into the chronic pain stage would have higherelevations on the negative cognitive and affective variables thanpatients whose pain was within the acute pain stage. Inaddition, it was expected that chronic congruent patients whotend to cope better with their pain than chronic incongruent painpatients, would only have slightly higher scores on the cognitiveand affective variables than acute pain patients. The acute painpatient's psychological reaction was still expected to besomewhat negative, however, since acute pain is assumed to beassociated with elevated levels of anxiety (Sternbach, 1985) andthe psychological reactions of acute and chronic pain patients(with no differentiation made between chronic congruent and120incongruent patients) have been found to be remarkably similar(Ackerman & Stevens, 1989; Philips & Grant, 1991 a).When the present findings concerning group differences oncognitive and affective variables are considered in the contextof other group differences on demographic and patient painrelated variables two alternative interpretations of the resultsare plausible. One interpretation suggests that in the acutepain stage, patients develop a negative psychological reaction totheir pain (e.g., respond emotionally, catastrophize), and tendto use passive strategies to cope (e.g., praying and hoping, useof opiate analgesics, decreased physical activity). In addition,since acute pain is common among laborers it is likely thatpatients are likely to be from a lower socioeconomic status, andto be on compensation since it is difficult for them to return tomore demanding work. The onset of pain is likely to result ingreater stress and uncertainty for some patients than others, andas a result some patients may begin to show incongruent signs andsymptoms because of their excessive focus of attention on theirpain or in attempt to gain the attention of their health careproviders and have their pain condition taken seriously. Thisinterpretation of pain also suggests that as pain persists, as itdoes for 15-40% of acute pain patients (Philips & Grant, 1991 b)some patients will adapt, while others may persist in their acutepain reaction. Those patients who adapt are likely to be underless stress (e.g., from a higher socioeconomic status and not oncompensation), and will show a lessened psychological reaction,and fewer attempts to cope passively with the pain. Because they121are not under as much stress and are coping more effectively withtheir pain these patients are not likely to show incongruentsymptoms and signs. Patients who do not adapt, on the otherhand, will persist in the acute pain reaction (e.g., negativepsychological reaction, and passive attempts to cope), andincongruent symptoms will become more frequent as the patient'sbecome excessively focused on their pain and attempt to gain theattention of their health care providers. The continual use ofpassive coping strategies (e.g., inactivity) may in turn resultin increased physical impairment in those who do not adapt ascompared to those who do. In general, this interpretation ofincongruent signs would suggest that incongruent signs andsymptoms covary with stress, and that patients who showincongruent signs and symptoms in the acute pain stage are notnecessarily the same patients who show incongruent signs in thechronic pain stage.An alternative interpretation of the results is thatpatients are predisposed through learning, environmental, and orbiological factors to display pain that is either incongruent orcongruent. Incongruent pain is used loosely here to refer topatients who are predisposed to have a negative psychologicalreaction, cope passively, focus excessively on their somaticsymptoms, and display and report signs and symptoms that areincongruent with underlying anatomy and physiology. Congruentpain on the other hand involves a lesser psychological reactionto the pain, fewer attempts to cope passively, less somaticfocusing, and few or no incongruent signs and symptoms of pain.122One factor which may be related to the predisposition of patientsto have incongruent pain may be socioeconomic status withpatients from a lower socioeconomic status more likely to displayincongruent pain than patients from a higher socioeconomicstatus. This overall interpretation of pain suggests that thecongruent and incongruent pain patterns will be present in theacute pain stage, and will remain relatively stable if painpersists into the chronic pain stage. It may also suggest thatmore patients with incongruent pain will develop acute LBP tobegin with, and be more likely to have their pain persist intothe chronic pain stage.Once again future research is needed to address whetherpatients are predisposed (perhaps because of biologicaldifferences or because of differences in learning histories) todisplay incongruent pain, or if these symptoms occur as a resultof stress. For instance, if a longitudinal study of LBP showedthat incongruent signs varied as a function of stress, supportfor a stress model of incongruent pain would be provided. If onthe other hand, a longitudinal study of LBP showed that patientswho have incongruent signs and symptoms in the acute pain stagecontinue to have incongruent signs and symptoms in the chronicpain stage, and the same was found of congruent pain patientssupport for a dispositional model of incongruent pain would beprovided.If support for a stress model of incongruent pain were foundfuture research would need to address the factors which are mostimportant in creating this reaction (e.g., physical impairment,123off from work, from a lower socioeconomic status, inadequatesocial support). On the other hand if support for adispositional model of incongruent pain were found, one wouldneed to examine what types of variables are important in creatingthe incongruent and congruent pain patterns. In general, futureresearch needs to examine how effective the current methods ofidentifying incongruent pain are with acute pain patients (e.g.,how many incongruent signs should be present before oneidentifies an acute pain patient as having incongruent pain).Future research needs to also address whether incongruent painpatterns are amenable to psychological treatment, and whethertreatment is more effective if patients with incongruent painpatterns are identified earlier in the pain cycle. It may bethat the incongruent patients require a more psychological basedtreatment package with graded and gradual rehabilitation back tonormal functioning, whereas congruent patients may be able to betreated effectively with a more traditional approach (e.g., rest,physiotherapy, medication) with only minimal psychologicalcounselling.If nothing else the above results concerning groupdifferences on the cognitive and affective have importantimplications for the treatment of acute pain patients in general.Traditionally, several assumptions about acute pain have beenmade which have affected treatment. First, it has been assumedthat acute pain although related to cognitive and affectivevariables (e.g., anxiety) is more related to tissue pathology(Sternbach, 1985). The results of the present study, and those124of others (Ackerman & Stevens, 1989; Philips & Grant, 1991 a)suggest that cognitive and affective variables play a much largerrole in acute pain, and in fact parallel the role played in themost distressed chronic pain patients. Second, it has beenassumed that the distress of acute pain patients is transient andfor only a small minority (10-15%) of patients will the distresscontinue into the chronic pain stage (Nachemson, 1982). A recentstudy by Philips and Grant (1991 b), however, suggests that thedistress of the acute pain patient continues over a period of atleast one and half months, and that for a large proportion ofpatients (40%) the pain persists into the chronic pain stage (6months). The consequence of these assumptions, which to thispoint have received no support has been that psychologicalinterventions for the treatment of acute LBP have largely beenignored, and the use of physical and pharmaceutical interventionsto relieve pain have been focused upon (Philips & Grant, 1991 c).If future research continues to find that ineffective copingstrategies, emotional reactions to pain, and dysfunctionalcognitions are characteristic of the acute pain patient (whetherthey are showing signs of incongruent pain or not) psychologicalinterventions at early stages will most certainly be implicated.Verbal Report of Pain. In the present study, cognitive andaffective features of the pain condition were expected to mediatethe verbal report of pain. Therefore, it was hypothesized thatacute pain patients, and to a lesser extent chronic congruentpain patients, who were predicted to have somewhat lower scoreson the cognitive and affective variables, would report less1.25intense and unpleasant pain than chronic incongruent patients.Because of the longer duration of the pain condition, chroniccongruent patients were expected to report more intense andunpleasant pain than acute pain patients, although thesedifferences were expected to be minimal. No support for thesehypotheses were found.Instead we found no group differences in the intensity ofpain in response to the leg movement or in response to the painexperienced on a typical day. On average all groups reportedmore intense pain on a daily basis than they did in response tothe leg movements. The average pain experienced on a daily basisroughly corresponded to the verbal descriptor strong, while thepain in response to the leg movements roughly corresponded to theverbal descriptor slightly intense.While no group differences were found in the intensity ofpain, group differences were found on the unpleasantness of painboth in response to the leg movement and in response to the painexperienced on a typical day. Acute and chronic incongruentpatients were found to report their pain to be similar in itsunpleasantness, but only chronic incongruent pain patientsdiffered significantly from chronic congruent patients inreported pain unpleasantness scores.In general, all groups were found to report the pain theyexperienced on a daily basis to be more unpleasant than the painthey experienced in response to the leg movements. The meanunpleasantness of pain experienced on a typical day roughlycorresponded to the verbal descriptor "very distressing", whereas1.26the unpleasantness of the pain in response to the leg movementsroughly corresponded to the verbal descriptor "very unpleasant".Facial Expressions of Pain. Cognitive and affectivevariables were also expected to mediate the expression of pain.Therefore, it was hypothesized that acute and to a lesser extentchronic congruent patients, who were also predicted to have lowerscores on the negative cognitive and affective variables, wouldbe less nonverbally expressive of their pain than chronicincongruent patients. Chronic congruent patients, because of theduration of their pain complaint, were expected to show more painrelated facial activity than acute pain patients. No support forthe hypotheses were found.No group differences were found in either the frequency orintensity of action units that occurred when subjects were askedto mask, genuinely express or exaggerate pain. Finding nodifferences between chronic congruent and chronic incongruentpatients was particularly surprising since chronic incongruentpatients are often assumed to be exaggerating or overreacting totheir pain. This assumption is so prevalent that one sign ofincongruent pain is the overreaction to pain upon examination(Waddell et al., 1980).The lack of findings may be in part be attributed to theprocedures used to study facial expressions. First, using theseprocedures the subjects may not have been motivated to exaggerateor overreact to their pain. Second, since subjects were veryaware of the fact that their facial actions were being videotapedand studied, they may have responded differently than if they had127not been aware. Third, having subjects mask, genuinely expressand exaggerate facial expressions of pain may have resulted indifferent expressions of pain than if subjects had not been askedto carry out these instructions. To adequately test thehypothesis set out in this study future research should perhapsutilize a between subjects design.Discrimination Among Pain Groups. It should be recalledthat the groups were found to differ with respect to a number ofimportant demographic and patient pain related variables.Ideally all of these variables would have been used as covariatesin the analyses of the dependent variables. However, twovariables were unsuitable as covariates (i.e., reported use ofpain killers, and disability) because of their large degree ofmeasurement error and subjectivity, one variable (i.e.,compensation status) was unsuitable as a covariate because of itsdichotomous nature, one variable (i.e., physical impairment) wasunsuitable because it was biased toward acute pain patientsobtaining lower scores than chronic pain patients, and onevariable (i.e., socioeconomic status) although suitable as acovariate, did not relate to the dependent variables of interest.Because of the difficulties in carrying out covarianceanalysis, the discriminant function analysis was of particularinterest. It allowed us to determine which variables, amongthose noted above which were significantly related to groupstatus, were the most important in discriminating among thegroups. When this analysis was carried out it was found that twosignificant discriminant functions emerged.12E'5The first discriminant function accounted for 49% of thevariability among groups, and appeared to maximally separateacute pain patients from the two chronic pain groups. Theloading matrix of correlations between predictors and the firstdiscriminant function suggested the best predictor fordistinguishing between the acute and the two chronic pain groupswas physical impairment. This was not surprising since themeasure of physical impairment used in this study was in partdetermined by the duration of the pain complaint, and the numberof surgeries, both of which are known to be greater for chronicas compared to acute pain patients.The second discriminant function was also significant andaccounted for 23% of the between group variability. Thisfunction maximally separated chronic incongruent pain patientsfrom chronic congruent pain patients, with acute pain patients inbetween the two. This function was correlated with an array ofcharacteristics the most important of which included: 1)emotional responses to pain; 2) presence of catastrophizingcognitions; 3) reported unpleasantness of the leg movement; 4)physical impairment; and 5) socioeconomic status. The results,suggested that chronic incongruent patients can be characterizedby their greater negative subjective interpretation of their pain(e.g., emotional responses to pain, catastrophizing cognitions,perceived unpleasantness of pain) greater physical impairment,and lower socioeconomic status compared to chronic congruent painpatients.129In general, consideration of both discriminant functionsresulted in the correct classification of the groups 71% of thetime. More specifically, acute pain patients were correctlyclassified as acute 83% of the time, and incorrectly classifiedas chronic congruent 17% of the time. Chronic congruent patientson the other hand were correctly classified as congruent 67% ofthe time, and incorrectly classified as acute 13% of the time,and incongruent 20% of the time. Finally, chronic incongruentsubjects were correctly classified as incongruent 63% of thetime, and incorrectly classified as acute 10% of the time, andcongruent 27% of the time. While these results are encouragingthere is definitely much room for improvement. Future researchcan perhaps be directed toward finding other important variableswhich may help in the classification of the pain groups.Discrimination Amona Chronic Pain Groups. Since previousresearch has primarily focused on differences among chroniccongruent and incongruent pain patients it was decided to carryout a further discriminant function analysis using the variablesto discriminate only between the congruent and incongruent paingroups. This analysis also seemed in order since in practice itis easy to distinguish between acute and chronic pain patients onthe basis of duration of pain complaint, and difficult todistinguish between chronic pain groups with varying degrees ofincongruent pain presentation.In doing this it was found that the discriminant functionaccounted for 32% of the between group variability suggestingthat much variability between groups was left unexplained. Theloading matrix of correlations between the predictors and thediscriminant function suggested that the best predictors of groupstatus were: 1) variables related to the negative interpretationof pain (i.e., emotionality, catastrophizing, movementunpleasantness); 2) compensation status; and 3) reported regularuse of pain killers.^That is, chronic incongruent patientscould be discriminated from chronic congruent patients in termsof their increased likelihood of negatively interpreting theirpain, being on compensation, and using pain killers on a regularbasis. Consideration of the discriminant function resulted incorrect classification of the groups 72% of the time. Morespecifically, it was found that congruent pain patients werecorrectly classified 80% of the time, while incongruent patientswere correctly classified 63% of the time.The results of the discriminant function analysis aresomewhat different from the results of the discriminant functionanalysis reported in the study of chronic congruent andincongruent LBP carried out by Reesor and Craig (1988).^Intheir study they similarly found catastrophizing to be animportant variable in discriminating among groups. In addition,however, they also found physical impairment, sense of controland affective pain ratings to be important as well. Theirclassification results were also somewhat better which may be inpart a result of a larger sample (40 congruent and 40incongruent) patients. More specifically, they found that 81% oftheir patients were correctly classified.131Differences in the importance of physical impairment indiscriminating among groups is somewhat troubling, and it restswith future studies to determine whether the differences are aneffect of sampling from different patient populations (e.g.,Reesor and Craig's patients were sampled from a tertiary carecenter, whereas ours were sampled from a primary care, moreheterogeneous community based sample). Regardless of differencesin patient populations, however, the importance ofcatastrophizing in discriminating among congruent and incongruentpain groups is clear, and strongly suggests that chronicincongruent patients are having difficulty coping.Facial Expressions of PainGenuine, Masked and Exaggerated Expressions. It should berecalled that on the basis of previous research it washypothesized that subjects, regardless of group status, wouldshow a consistent set of facial actions associated with pain,and, in addition, would be remarkably successful in both maskingand exaggerating their pain. These hypotheses were largelysupported. With regard to AU frequency, brow lowering (AU 4),orbit tightening (AUs 6 and 7), levator contraction (AUs 9 and10), and mouth opening (AUs 25, 26 and 27) were all found tooccur more frequently during the genuine pain segment than duringthe baseline segment when subjects were simply instructed tocarry out the non-painful task of wiggling their toes. Examiningthe intensity of AUs was a less useful way of discriminatingamong the genuine and baseline segments. It was found, forinstance, that only orbit tightening (AUs 6 and 7), lip press (AU13.224), and mouth opening (AUs 25, 26 and 27) were more intenseduring the genuine pain segment than during the baseline segment.The above results are comparable to those that have beenfound in studies of facial expressions of clinical andexperimental pain. That is the AUs which were found in thepresent study to occur more frequently during the genuine painsegment rather than the baseline segment have also been found tobe consistently related to pain in other studies (Craig, Hyde &Patrick, 1991). The most relevant study to compare the findingsto is the study carried out by Craig, Hyde & Patrick (1991) inwhich facial expressions of LBP patents in response tophysiotherapy leg movements were studied.^Although they used aslightly different method for coding the facial actions they alsofound brow lowering (AU 4), cheek raising (AU 6), tightening ofthe eye lids (AU 7), raising of the upper lip (AU 10), andparting of the lips (AU 25) to occur more frequently in responseto the painful leg movements, than in response to baseline. Inaddition, however, they found eye closing (AU 43) to be morefrequently associated with pain; this was not found in thepresent study.With regard to the subject's ability to mask pain,comparisons between baseline, masked, and genuine segments aremost relevant. That is, subjects would be viewed as having beensuccessful in masking their pain if their facial expressionsduring the masked pain segment were similar to those found duringbaseline periods, and dissimilar from those found during genuinepain segments. Partial support for the subject's ability to maskpain was found in that the frequency and intensity of AUs foundduring baseline and masked pain segments were not significantlydifferent. However, the frequency and intensity of AUs foundduring the genuine pain segment and the masked pain segment werealso not significantly different except with respect to mouthopening which was found to be slightly more frequent and intenseduring the genuine pain segment, as compared to the masked painsegment. Therefore, while subjects were successful in reducingtheir facial activity during the masked pain segment to a pointwhere the activity was not significantly different from baseline,it was not reduced enough to result in significant differencesbetween genuine and masked pain segments. This suggests thatwhile residual cues of pain may be present on the face it islikely extremely difficult to determine whether the patient is inpain.The above results are somewhat comparable to the resultsfound by Craig, Hyde, and Patrick (1991) who also found fewdifferences between masked and baseline segments. The exceptionwas: 1) AU 7 (lid tightening) which was found to be marginallymore frequent during the masked segment as compared to thebaseline segment; and 2) AU 45 (blinking) which was found to besignificantly less frequent during the masked segment compared tothe baseline segment. With regard to differences between maskedand genuine segments, outer brow raise (AU 2), cheek raise (AU6), parted lips (AU 25), closed eyes (AU 43) and blinking (AU 45)were found to be less frequent during the masked pain segment ascompared to the genuine pain segment. Similar to the presentstudy they found no differences between the genuine and maskedpain segments with respect to several important pain AUs, namelybrow lowering (AU 4), lid tightening (AU 7), levator contraction(AUs 9 and 10), jaw dropping or mouth stretch (AUs 26 and 27).Together the results suggest that, while subjects instructed tomask expressions of pain in both studies decreased their facialactivity to baseline levels for the most part, many facialactions continued to occur at levels which were comparable tothose expected when subjects were genuinely expressing pain.With regard to the exaggerated expressions of pain thecomparison that was of primary interest was that between genuineand exaggerated pain segments. This comparison showed thatseveral AUs which were associated with pain in the present study,namely brow lowering (AU 4), and orbit tightening (AUs 6 and 7),were more frequent during the exaggerated pain segment thanduring the genuine pain segment. In addition, two AUs which werenot associated with pain in the present study, but have beenassociated with pain in previous studies (e.g., lip stretch (AU20) and eyes closed (AU 43) were also more frequent. With regardto AU intensity several pain associated AUs (brow lowering (AU4), orbit tightening (AU 6 and 7), and levator contraction (AUs 9and 10)) increased in intensity, while several other non-painrelated AUs (i.e., inner and outer brow raise (AUs 1 and 2), lipstretch (AU 20) and eye closing (AU 43) also increased inintensity as well. These results suggest that when subjects wereexaggerating their pain they were able to keep much of the samefacial expression that would be expected when subjects were135genuinely expressing pain. In addition, however, they addedconsiderably to the expressionIt is difficult to compare the above results to those foundby Craig, Hyde and Patrick (1991) since the instructions in theirstudy were significantly different from the instructions given tothe subjects in the present study. That is, in the present studysubjects were asked to exaggerate their expressions of pain inresponse to a painful movement, whereas in the study carried outby Craig, Hyde and Patrick (1991) subjects were asked to fake anexpression of pain in response to a non-painful movement. Ourchoice to have subjects exaggerate rather than fake was madebecause we felt that incongruent pain was likely to be morerelated to exaggerated expressions of pain, rather than fakedexpressions of pain. Nevertheless, some comparison can be madebetween the exaggerated expressions of pain in the present study,and the faked expressions of pain in the study by Craig andcolleagues (1991). Craig and colleagues, for instance, alsofound that brow lowering (Au 4) and cheek raising (AU 6)increased during the faked as compared to the genuine painsegment. In addition, however, they also found that pulling atthe corner of the lips (AU 12) increased and blinking (AU 45)decreased. The similarities in results may suggest that there isa relationship between exaggerated and faked expressions of pain.Relations Among Genuine Facial Actions of Pain. From theabove comparison it was clear that there were a number of AUswhich were related to pain, but it was not clear how the AUsrelated to each other. It was decided, therefore, to carry out aprincipal components analysis of the AUs to determine how theyinterrelated empirically. When this was done it was found thatthe AUs were all highly correlated, and loaded on one single painfactor which accounted for approximately 44 percent of thevariance in facial activity in the genuine pain segment. Thelarge amount of variance which was left unexplained suggests thatthere is wide variation in the individual expression of pain.Despite this, it appears that brow lowering (AU 4), orbittightening (AUs 6 and 7), levator contraction (AUs 9 and 10), andmouth opening (AUs 25, 26 and 27) carry the bulk of theinformation about pain. These results compare to those reportedby Prkachin (1992) who studied facial actions which wereconsistently related to four experimental modalities of pain(e.g., cold pressor, ischemia, electric shock, and mechanicalpressure). He found brow lowering (AU 4), orbit tightening (AUs6 and 7), levator contraction (AUs 9 and 10) and eyes closing (AU43) to carry the bulk of the pain information. The onlydifferences between the two studies rests with the importance ofmouth opening in the present study, and the importance of eyeclosing in Prkachin's study. These could be a result ofdifferences in the intensity and nature of pain experiencedbetween the two studies.It has been proposed, for instance, that more facial actionsoccur as pain becomes more severe and enduring (Prkachin &Mercer, 1989). At low levels of pain, facial actions that areconsistently related to pain (e.g., brow lowering, and orbittightening) are expected to be observed. At moderate levels of137pain, middle facial actions (e.g., levator contraction) areexpected to occur in addition to those already identified.Finally, with the most intense levels of pain, lower facialactions (e.g., mouth opening, horizontal stretch mouth) areexpected to come into play.If we interpret the results of the present study, and theresults of Prkachin's study in line with this conceptualization,it is likely that the pain experienced in response to the legmovements was of a more intense nature than the pain experiencedin response to the experimental pain in Prkachin's study. That, in the present study mouth opening was found to beimportantly related to pain, whereas in the study conducted byPrkachin it was not. Despite differences in results, what isimportant to emphasize about these results is that in bothstudies the experience of pain was found to be consistentlyrelated to a number of facial actions which suggests that thepresence of genuine pain in patients should be easy to identify.Predictors of Verbal and Nonverbal Measures of PainSince the verbal and nonverbal measures of pain weregenerally of limited use in discriminating among groups it was ofinterest to examine which of all of the variables measured in thestudy could be used to predict the verbal and nonverbal measuresof pain. The variables in the study were divided into sixdistinct groups (see Results). Variables from each group wereused to predict pain intensity, pain unpleasantness, and genuinepainful facial activity, and then those variables which weresignificant within their groups were entered together into afinal stepwise regression analysis.Predictors of Movement Intensity. With regard to painintensity, several variables (e.g., the pain drawing, thenonorganic signs assessment, first language spoken, socioeconomicstatus, regular use of opiate analgesics, tendency tocatastrophize, and to use behavioural coping strategies,unpleasantness of the movement, intensity of pain experienced ona daily basis, and the genuine painful facial activity score)were found to predict the intensity of the pain when they wereentered into a regression analysis that only included othersimilar variables as predictors. When all of these significantpredictors were entered together into a regression analysis,however, variables which appeared to be importantly related tointensity of pain were found to share variance with moreimportant variables, and thus failed to make a uniquecontribution. The results of this analysis suggested that theverbal report measures of pain were all highly related since themost important variables in predicting pain intensity were foundto be other verbal report measures of pain (e.g., movementunpleasantness, and daily pain intensity).The results also suggested that the use of behaviouralstrategies to cope with pain. The unique contribution of thebehavioural strategies to the prediction of pain intensity was atfirst somewhat surprising since one current and apparentlyeffective approach to treating pain is to encourage normalactivity as much as possible irrespective of pain (Philips E.13•4Grant, 1991 c). These results at first would seem to suggestthat such an approach is not effective in reducing pain intensitysince patients who cope with their pain more actively (e.g.,going to movies, shopping, doing projects, household chores) alsoreported more intense pain. However, the approach that isdescribed above also typically involves graded exposure such thatpatients begin their return to activity at low levels and buildup. Although it is difficult to determine, it may have been thatthe patients in this sample did not grade their exposure, butrather immediately returned to normal activity. In general, theresults suggest the relation between activity (e.g., normalversus graded exposure) and pain intensity needs to be furtherexplored.Predictors of Pain Unpleasantness. With respect to theunpleasantness of pain, several variables (e.g., the nonorganicsigns assessment, the pain drawing, compensation status, the useof anti-inflammatories on a regular basis, degree of worry, theintensity of the movement, and the facial activity found duringthe genuine pain segment) were found to predict the rating whenthey were entered into the regression analyses with other similarpredictor variables. When they were all entered together,however, the most important variable in predicting theunpleasantness of the pain was the reported intensity of thepain. Together, the above results suggest that our hypothesesthat cognitive and affective features of pain would mediate theverbal report of pain were not supported.140Predictors of Genuine Painful Facial Activity. Severalvariables (e.g., the pain drawing, the subject's marital status,the reported intensity of the movement, and the overall facialactivity of subjects during the exaggerated segment and thebaseline segment) were found to predict painful facial activitywhen entered into a regression analysis that included variablesonly from that group. When these significant predictors wereentered together, however, the degree to which subjects expressedpain in the genuine pain segment was largely a function of howmuch they were willing to exaggerate their pain, and how muchfacial activity they showed during a baseline segment. Inaddition, the genuine facial expression of pain was also relatedto how unpleasant subjects found the painful leg movement. Theseresults suggest first, that there is a small, but neverthelesspositive correlation between the nonverbal and verbal expressionsof pain. Second, they suggest that the degree to which subjectsexpress pain is largely a function of how expressive they are inthe first place (baseline facial activity) or how willing theyare to actively exaggerate pain. The implication is that if onewishes to get an accurate assessment of the patient's nonverbalexpression of pain, one must take into account how expressivethat person is to begin with.Summary and Directions for Future ResearchThe purpose of the present study was to examine how thepsychological reactions to pain of acute LBP patients wouldcompare to the psychological reactions to pain of chronic LBPpatients who displayed signs and symptoms which were congruent or141incongruent with underlying anatomy and physiology. It was alsoof interest to examine whether verbal and nonverbal measures ofpain could be used to discriminate among pain groups.In the literature describing acute and chronic painconditions it is often assumed that as pain persists thepsychological distress of the patient will increase (Sternbach,1985). This led us to hypothesize that acute pain patients wouldobtain lower scores on cognitive and affective measures of painthan chronic pain patients. Research also suggests, however,that there is wide variability in the psychological reactions ofchronic pain patients, with some chronic pain patients (e.g.,chronic congruent patients) coping better with their pain thanothers (e.g., chronic incongruent patients). This led us tofurther hypothesize that the differences between acute painpatients and chronic congruent pain patients on the cognitive andaffective variables would be much smaller than the differencesbetween acute and chronic incongruent pain patients. We furtherhypothesized that cognitive and affective variables would mediatethe verbal report of pain and the nonverbal expression of pain,such that acute pain patients and chronic congruent painpatients, who were predicted to have lower scores on the negativecognitive and affective measures, would also be less verbally andnonverbally expressive of their pain than chronic incongruentpain patients.Contrary to our hypotheses quite a different pattern ofresults emerged. To begin we found that acute and chronicincongruent pain patients responded more emotionally to theirpain, were more likely to use passive coping strategies, and morefrequently reported dysfunctional, catastrophizing cognitionsthan chronic congruent pain patients. In addition to having asimilar negative psychological reaction to pain, acute andchronic incongruent patients reported their pain to be similar inits unpleasantness, although only the chronic incongruentpatients differed significantly from the chronic congruentpatients. Acute and chronic incongruent patients were also morelikely to be from a lower socioeconomic status, on compensation,using opiate analgesics on a regular basis and reporting greaterdisability as a result of the pain than chronic congruentpatients. Groups were also found to differ with respect tophysical impairment with chronic incongruent patients scoringhighest on the measure followed by chronic congruent and acutepain patients. All differences were significant.In carrying out several discriminant function analyses itwas found that acute pain patients could be maximally separatedfrom the two chronic pain groups on the physical impairmentmeasure. This analysis was largely biased, however, since themeasure of physical impairment was biased toward acute painpatients obtaining lower scores on the measure than chronic painpatients. A further discriminant function analysis suggestedthat the chronic incongruent patients could be discriminatedfrom chronic congruent patients in terms of their increasedlikelihood of negatively interpreting their pain, being oncompensation, and using pain killers on a regular basis.The pattern of results that were found in the present studyare subject to two alternative interpretations. Oneinterpretation of the results is that whether patients displayincongruent signs and symptoms is largely a function of how muchstress patients are under, how uncertain they are about theirpain condition, and how difficult it is for them to cope withtheir pain. This interpretation further suggests that in theacute pain stage all patients will have a negative psychologicalreaction to pain, and those who are having the most difficultiescoping with their pain will show incongruent signs and symptoms.Further as pain persists some patients will adapt, while othersmay persist in their acute pain reaction. Those patients whoadapt are likely to be under less stress (e.g., from a highersocioeconomic status and not on compensation), and as a resultwill show a lessened psychological reaction, fewer attempts tocope passively with the pain, and as a result few incongruentsymptoms and signs. Patients who do not adapt, on the otherhand, will persist in the acute pain reaction (e.g., negativepsychological reaction, and passive attempts to cope), andincongruent symptoms will become more frequent as the patient'sbecome excessively focused on their pain and attempt to gain theattention of their health care providers in order that their paincondition be taken seriously. The continual use of passivecoping strategies (e.g., inactivity) may in turn result inincreased physical impairment in those who do not adapt ascompared to those who do. This model predicts that inocngurentsigns will covary with stress, and that patients who show144incongruent signs and symptoms at one stage need not showincongruent signs and symptoms at another.An alternative interpretation of the results is thatpatients are predisposed through learning, environmental, and orbiological factors to display pain that is either incongruent orcongruent. Incongruent pain is used loosely here to refer topatients who are predisposed to have a negative psychologicalreaction, cope passively, focus excessively on their somaticsymptoms, and display and report signs and symptoms that areincongruent with underlying anatomy and physiology. Congruentpain on the other hand involves a lesser psychological reactionto the pain, fewer attempts to cope passively, less somaticfocusing, and few or no incongruent signs and symptoms of pain.One factor which may be related to the predisposition of patientsto have incongruent pain may be coming from a lower socioeconomicstrata. This overall interpretation of pain suggests that thecongruent and incongruent pain patterns will be present in theacute pain stage, and will remain relatively stable if painpersists into the chronic pain stage. It may also suggest thatmore patients with incongruent pain will develop acute LBP tobegin with, and be more likely to have their pain persist intothe chronic pain stage.Future research is needed to address whether patients arepredisposed (perhaps because of biological differences or becauseof differences in learning histories) to display incongruent painor if incongruent symptoms occur as a result of stress. Futureresearch should also examine whether different types of treatment1145would be more effective with congruent as compared to incongruentpatients.In general, the results showing that acute pain patients doindeed have a strong psychological reaction to pain haveimportant implications for the treatment of acute pain patientsin general. That is, if future research continues to find thatineffective coping strategies, emotional reactions to pain, anddysfunctional cognitions are characteristic of the acute painpatient (whether they are showing signs of incongruent pain ornot) psychological interventions at early stages will mostcertainly be implicated, and in need of evaluation.One of the most surprising conclusions from the study wasthe fact that no differences in nonverbal expressions of painamong groups of chronic congruent and incongruent patients werefound. While this may indeed be the case, the lack of findingsmay in part be an artifact of the design. That is, differentresults may have been found if subjects were not aware that theywere being observed, and if subjects had not been asked to maskand exaggerate their pain. This procedure may have resulted inthe amplification or attenuation of the genuine pain expressionsthat would not normally have resulted if patients had not beenasked to follow these instructions. Future research which eitheronly asks patients to genuinely express their pain, or uses abetween groups design to study masked, exaggerated and genuineexpressions of pain would better be able to test hypothesesconcerning facial expressions of pain.146While no group or sex differences were found with respect toexpressions of pain, differences in facial activity did emerge asa result of the instructional sets. The facial actions whichwere associated with pain in the present study were highly inter-related, and were consistent with facial expressions which werefound to be associated with pain in past research. The facialactions associated with pain included: brow lowering, orbittightening, levator contraction, and mouth opening. Whensubjects were asked to mask their facial expressions of pain theywere successful in the sense that they reduced their facialactions to a level expected only to be found during baseline.They were unsuccessful in the sense that their facial activityduring the masked pain segment, although reduced, still did notdiffer significantly from the genuine pain segment. The exceptionto this was the degree of mouth opening which was found to besignificantly less during both baseline, and masked pain segmentscompared to the genuine pain segment. It is interesting to notethat mouth opening is expected to occur with only the mostintense levels of pain (Prkachin & Mercer, 1989), and thatsubjects in effect unknowingly reduced this action when maskingpain.With regard to exaggerated expressions of pain, once againsubjects were remarkably successful in following theinstructions. That is, they tended to increase the frequency andintensity of all facial actions that were associated with pain.In addition, however, subjects also added and increased theintensity of several non-pain related actions. Using these non-4 7pain related actions as cues it may be possible to detectpatients who are indeed exaggerating pain.The final point that needs to be made, is that in thepresent study assumed differences between acute and chronic LBPpatients, led us to make certain hypotheses about how thesepatient groups would cope with and express their pain. Thequestion that this raises is whether health care professionalsalso make these assumptions in their assessment of pain. Futureresearch needs to be directed toward understanding theassumptions that health care professionals make, how theseassumptions in turn effect diagnosis and treatment of patients,and how in general health care professionals use the vast arrayof data (e.g., physiological, psychological, verbal and nonverbalexpressions) that is available to them when they are makingjudgements about another's pain.148REFERENCESAckerman, M. D. & Stevens, M. (1989). Acute and chronic pain:Pain dimensions and psychological status. Journal of Clinical Psychology, 45 (2), 223-228.Beals, R. K. & Hickman, N. W. (1972). Industrial injuries of theback and extremities: Comprehensive evaluation - an aid inprognosis and management: A study of one hundred and eightypatients. Journal of Bone and Joint Surgery, 54, 1593-1611.Beecher, H. K. (1955). The powerful placebo. Journal of theAmerican Medical Association, 159, 1602.Bellissimo, A. & Tunks, E. (1984). Chronic pain: Thepsychotherapeutic spectrum. New York: Praeger.Biedermann, H. J., Monga, T. N., Shanks, G. L. & McGhie, A.(1986). The classification of back pain patients:Functional versus organic. Journal of Psychosomatic Research, 30, 273-276.Bigos, S. J., Crites, B. M. (1987). Acute care to prevent backdisability. Clinical Orthopedics, 221, 121-130.Blishen, B. R., Carroll, W. K. & Moore, C. (1987). The 1981socioeconomic index for occupations in Canada. CanadianReview of Sociology and Anthropology, 24, 465-488.Block, A. R., Kremer, E. & Gaylor, M. (1980). Behaviouraltreatment of chronic pain: The spouse as a discriminativecue for pain behavior. Pain, 9, 243-252.Blumetti, A. E. & Modesti, L. M. (1976). Psychologicalpredictors of success or failure of surgical interventionfor intractable back pain. In J. J. Bonica & D. Able-Fessard (Eds.), Advances in pain research and therapy. NewYork: Raven Press.Bond, M. R. (1979). Pain: Its nature, analysis and treatment.London: Churchill Livingston.Bonica, J. J. (1980). Pain research and therapy: Past andcurrent status and future needs. In L. K. Y. Ng & J. J.Bonica (Eds.), Pain discomfort and humanitarian care.Amsterdam: Elsevier/ North-Holland.Brena, S. F. & Chapman, S. L. (1981). The "learned painsyndrome" decoding a patient's signals. Postgraduate Medicine, 69 (1), 53-62.Chambers, W. G. & Price, G. G. (1967). Influence of nurse uponeffects of analgesics administered. Nursing Research,228-233.149Chapman, C. R. (1977). Psychological aspects of pain patienttreatment. Archives of Surgery, 112, 767-772.Craig, K. D., (1984). Emotional aspects of pain. In R. Melzack& P. D. Wall (Eds.), Textbook of pain. Edinburgh:Churchill/Livingston.Craig, K. D., Best, H. & Reith, G. (1974). Social determinantsof reports of pain in the absence of painful stimulation.Canadian Journal of Behavioural Sciences, 6 (2) 169-177.Craig, K. D., Hyde, S. & Patrick, C. J. (1991). Genuine,suppressed and faked facial behavior during exacerbation ofchronic low back pain. Pain,Craig, K. D. & Patrick, C. (1985). Facial expression duringinduced pain. Journal of Personality and Social Psychology,48, 1080-1091.Craig, K. D. & Prkachin, K. M. (1983). Nonverbal measures ofpain. In R. Melzack (Ed.), Pain measurement andassessment. New York: Raven Press.Craig, K. D., Prkachin, K. M. & Grunau, R. V. E. (In press). Thefacial expression of pain. In D. C. Turk & R. Melzack(Eds.), The handbook of pain assessment. New York: GuilfordPress.Craig, K. D. & Weiss, S. M. (1971). Vicarious influences onpain-threshold determinations. Journal of Personality andSocial Psychology, 11, 53-59.Crook, J., Rideout, E. & Browne, G. (1984). The prevalence ofpain complaints in a general population. Pain, 18, 299-314.Darwin, C. (1965). The expression of the emotions in man andanimals. Chicago: University of Chicago Press (originallypublished in 1872).DePaulo, B. M., Rosenthal, R., Eisenstat, R. A. Rogers, P. C. &Finkelstein, S. (1978). Decoding discrepant nonverbal cues.Journal of Personality and Social Psychology, 36, 313-323.Dilane, J. B., Fry, J. & Kalton, G. (1966). Acute back syndrome-a study from general practice. British Medical Journal, 2,82-84.Donham, G. W., Mikhail, S. F. & Meyers, R. (1984). Value ofconsensual ratings in differentiating organic and functionallow back pain patients. Journal of Clinical Psychology, 40(2), 432-439.150Doxey, N. C., Dzioba, R. B., Mitson, G. L. & Lacroix, J. M.(1988). Predictors of outcome in back surgery candidates.Journal of Clinical Psychology, AA (4), 611-622.Dudley, S. R. & Holm, K. (1984). Assessment of pain in relationto select nurse characteristics. Pain, 18 (2), 179-186.Duncan, G. H., Bushnell, M. C. & Lavigne, G. J.^(1989).Comparison of verbal and visual analogue scales formeasuring the intensity and unpleasantness of experimentalpain. Pain, 37, 295-303.Dzioba, R. B. & Doxey, N. C. (1984). A prospective investigationinto the orthopaedic and psychological predictors of outcomeof first lumbar surgery following industrial injury. Spine,9, 614-623.Ekman, P. (1977). Biological and cultural contributions to bodyand facial movement. In J. Blacking (Ed.), Anthropology of the body. London: Academic Press.Ekman, P. & Friesen, W. (1971). Constants across cultures in theface and emotion. Journal of Personality and Social Psychology, 17, 124-129.Ekman, P. & Friesen, W. (1978a). Investigator's guide to thefacial action coding system. Palo Alto, California:Consulting Psychologists Press.Ekman, P. & Friesen, W. (1978b). Manual for the facial actioncoding system. Palo Alto, California: ConsultingPsychologists Press.Ekman, P. & Friesen, W. (1983). EMFACS: Coders Instructions.Unpublished manuscript, University of California, HumanInteraction Laboratory, San Francisco, CA.Ekman, P. & Friesen, W. (1984). Unmasking the face. Palo Alto:Consulting Psychologists Press.Englebert, H. J. & Vrancken, M. A. E. (1984). Chronic pain fromthe perspective of health: A view based on systems theory.Social Science and Medicine, 19 (12), 1392-1984.Fairbank, J. C., Couper, J., Davies, J. B. & O'Brien, J.P.^(1980). The Oswestry Low Back Pain Questionnaire, Physiotherapy,66, 271-273.Finneson, B. E. (1976). Modulating effect of secondary gain onthe low back pain syndrome. In J. J. Bonica & D. Able-Fessard (Eds.), Advances in pain research and therapy. NewYork: Raven Press.151Fordyce, W. (1976). Behavioral methods for chronic pain andillness. St. Louis, Mo.: Mosby.Fordyce, W. (1978). Learning processes in pain. In R. A.Sternbach (Ed.), The psychology of pain. New York: RavenPressFordyce, W. (1983). The validity of pain behaviour measurement.In R. Melzack (Ed.) Pain measurement and assessment. NewYork: Raven Press.Fordyce, W. E., Brena, S. F., Holcomb, R. J., De Lateur, B. J. &Loeser, J. D. (1978). Relationship of patient semantic paindescriptors to physician diagnostic judgements, activitylevel measures and MMPI. Pain, 5, 293-303.Fordyce, W. E., Lansky, D., Calsyn, D. A., Shelton, J. L., Stolov, W. C. & Rock, D. L. (1984). Pain measurement and painbehaviour. Pain, 18, 53-69.Flor, H., Kerns, R. D. & Turk, D. C. (1987). The role of^thespouse in the maintenance of chronic pain, Journal of Psychosomatic Research, 31, 251-260.Freeman, C., Calsyn, D. & Louks, J. (1976). The use of theMinnesota Multiphasic Personality Inventory with low backpain patients. Journal of Clinical Psychology, 32 (2), 294-298.Frymoyer, J. W., Pope M. H., Costanza, M. C., Rosen, J. C.,Goggin, J. E. & Wilder, D. G. (1980). Epidemiologic studiesof low-back pain. Spine, 5, 419 423.Gaston-Johansson, F. (1984). Pain assessment: Differences inquality and intensity of the words pain, ache, and hurt.Pain, 20, 69-76.Gentry, W. D., Newman, M. C., Goldner, J. L. & von Baeyer, C.(1977). Relation between graduated spinal block techniqueand MMPI for diagnosis and prognosis of chronic low-backpain. Spine, 2 (3), 210-213.Gracely, R. H. (1979). Psychophysical assessment of human pain.In J. Bonica (Ed.), Advances in Pain Research and Therapy,3. New York: Raven Press.Gracely, R. H. (1983). Pain language and ideal pain assessment.In R. Melzack (Ed.), Pain measurement and assessment. NewYork: Raven Press.Gracely, R. H. & Dubner, R. (1981). Pain assessment in humans-Areply to Hall. Pain, 11, 109-120.152Gracely, R., Dubner, R. & McGrath, P. (1979). Narcotic^analgesia: Fentanyl reduces the intensity but not theunpleasantness of painful pulp sensations. Science, 203,1261-1263.Gracely, R. H., McGrath, P. & Dubner, R., (1978a). Ratio^scales of sensory and affective verbal pain descriptors. Pain,5, 5-18.Gracely, R., McGrath, P. & Dubner, R. (1978b). Validity andsensitivity of ratio scales of sensory and affective verbalpain descriptors: Manipulation of affect by diazepam. Pain,5, 19-29.Grunau, R. V. E., Johnston, C. C. 64 . Craig, K. D. (1990). Facialactivity and cry to invasive and noninvasive tactilestimulation in neonates. Pain, 42 (3), 295-305.Guest, G. H. & Drummond, P. D. (1992). Effect of compensation onemotional state and disability in chronic back pain. Pain,48, 125-130.Hadjistavropoulos, H. D., von Baeyer, C. & Ross, M. (1990). Arephysicians ratings of pain effected by patients' physicalattractiveness? Social Science and Medicine, 31, 69-72.Hall, W. (1981). On "ratio scales of sensory and affectiveverbal pain descriptors". Pain, 11, 101-108.Hanvik, J. L. (1951). MMPI profiles in patients with low-backpain. Journal of Consulting and Clinical PsychologY, 15,350-353.Hardy, J. D. (1956). The nature of pain. Journal of ChronicDiseases, 7, 22-51.Harper, R. G., Weins, A. N. & Matarazzo, J. D. (1978). Nonverbal communication.^New York: Wiley.Hilgard, E., (1969). Pain as a puzzle for psychology andphysiology. American Psychologist, 103-113.Holand, M. K. & Tarlow, G. (1972). Blinking and mental load.Psychological Reports, 323, 119-127.Holand, M. K. & Tarlow, G. (1975). Blinking and thinking.Perceptual and Motor Skills, 41, 403-406.House, A. E., House, B. J. & Campbell, M. B. (1981). Measures ofinterobserver agreement: calculating formulas anddistribution effects. Journal of Behavioural Assessment, 3,37-57.153Hrudey, W. P. (1991). Overdiagnosis and Overtreatment of Low Back Pain: Long Term Effects. Unpublished Manuscript.Workers' Compensation Board of British Columbia, Vancouver,B.C.Hunter, M. & Philips, C. (1981). The experience of headache- Anassessment of the qualities of tension headache pain. Pain,10, 209-220.International Association for the Study of Pain, Classificationof Chronic Pain. (1986). Pain, (Suppl. 3), S1-S226.Jacox, A. K. (1980). The assessment of pain. In L. Smith, H.Mersky, and S. Gross (Eds.), Pain: Meaning and management.New York: Springer.Jamison, R. N., Sbrocco, T. & Parris, W. C. V. (1989). Theinfluence of physical and psychosocial factors on accuracyof memory for pain in chronic pain patients. Pain, 37, 289-294.Johnson, M. (1977). Assessment of clinical pain. In A. K. Jacox(Ed.), Pain: A source book for nurses and other health care professionals. Boston: Little Brown & Co.Johnson, J., Kirchoff, K. & Endress, M. P. (1975). Alteringchildren's distress behavior during orthopedic cast removal.Nursing Research, 824 (6), 404-410Joukamaa, M. (1987). Psychological factors in low back pain.Annals of Clinical Research, 19, 129-134.Kahn, E. A. (1966). On facial expression. Clinical Neurosurgery, 12, 9-22.Keefe, F. J. & Block, A. R. (1982). Development of anobservation method for assessing pain behavior in chroniclow back pain patients. Behavior Therapy, 13, 363-375.Keefe, F. J., Bradley, L. A. & Crisson, J. E. (1990).Behavioural assessment of low back pain: Identification ofpain behavior subgroups, Pain, 40, 153-160.Keefe, F. J., Crisson, J., Urban, B. J. & Williams, D. A. (1990).Analyzing chronic low back pain: The relative contributionsof pain coping strategies. Pain, 40, 293-301.Keefe, F. J., & Gill, K. M. (1986). Behavioural concepts in theanalysis of chronic pain syndromes, Journal of Consulting and Clinical Psychology, 5A, 776-783.Kelsey, J. L. & White, A. A. (1980). III: Epidemiology and impactof low-back pain. Spine, 1, 133-142.154Kendall, P., Williams, L., Pechacek, T., Shisslak, C. & Herzoff,N. (1979). Cognitive-behavioral and patient educationinterventions in cardiac catheterization procedures: ThePalo-Alto medical psychology project. _ Journal of Consulting and Clinical Psychology, 47, 49-58.Kleck, R. E., Vaughan, R. C., Cartwright-Smith, J., Vaughan, K.B., Colby, C. F. & Lanzetta, J. T. (1976). Effects of beingobserved on expressive, subjective, and physiologicalresponses to painful stimuli. Journal of Personality andSocial Psychology, 34, 1211-1218.Kraut, R. E. (1978). Verbal and nonverbal cues in the perceptionof lying. Journal of Personality and Social Psychology, 33,354-370.Kraut, R. E. (1982). Social presence, facial feedback, andemotion. Journal of Personality and Social Psychology, 42,853-863.Kremer, E. & Atkinson, J. H. (1981). Pain measurement: Constructvalidity of affective dimension of the McGill PainQuestionnaire with chronic benign pain patients. Pain, 11,93-100.Kremer, E. & Atkinson, J. H. (1983). Pain language as a measureof affect in chronic pain patients. In R. Melzack (Ed.),Pain measurement and assessment. New York: Raven Press.Kremer, E. F., Block, A. & Atkinson, J. H. (1983). Assessment ofpain behavior: Factors that distort self-report. In R.Melzack (Ed.), Pain measurement and assessment. New York:Raven Press.Lanzetta, J. T., Cartwright-Smith J. & Kleck, R. E. (1976).Effects of nonverbal dissimulation on emotion. Journal of Personality and Social Psychology, 33, 354-370.Lawson, K., Reesor, K., Keefe, F. J. & Turner, J. (1990).Dimensions of pain-related cognitive coping: cross-validation of the factor structure of the Coping StrategyQuestionnaire. Pain,Lazarus, R. (1986). Coping strategies. In S. McHugh & T. M.Vallis, (Eds.), Illness behaviour: A multidisciplinarymodel. New York: Plenum Press.Leavitt, F. (1985). The value of the MMPI conversion 'V' in theassessment of psychogenic pain. Journal of Psychosomatic Research, 29, 125-131.Leavitt, F., Garron, D.C. & Bieliauskas, L. A. (1980).Psychological disturbance and life event differences among155patients with low back pain. Journal of Consulting andClinical Psychology, 48, 115-116.Leavitt, F., Garron, D.C., D'Angelo, C. M. & McNeill, T. W.(1979). Low back pain in patients with and withoutdemonstrable organic disease. Pain, 6, 191-200.Lehmann, T. R., Russell, D. W. & Spratt, K. F. (1983). Theimpact of patients with nonorganic physical findings on acontrolled trial of transcutaneous electrical nervestimulation and electroacupuncture. Spine, 8, 625-634.Lenburg, C. B., Glass, H. P., & Davitz, L. J. (1970). Inferencesof physical pain and psychological distress in relation tothe stage of the patients' illness and occupation of theperceiver. Nursing Research, 19, 392-398.LeResche, L. (1982). Facial expression in pain: A study ofcandid photographs. Journal of Nonverbal Behaviour, 7, 464.LeResche, L. & Dwrokin, S. (1988). Facial expressions of painand emotion in chronic TMD patients, Pain, 35, 71-78.LeResche, L., Ehrlich, K. J. & Dwrokin, S. (1990). Facialexpressions of pain and masking smiles: Is "Grin and Bearit" a pain behaviour? Pain, (Suppl. 5), S286.Loeser, J. D. (1980). Low back pain. In J. J. Bonica (Ed.)Pain. New York: Raven Press.McCreary, C., Turner, J. & Dawson, E. (1979). The MMPI as apredictor of response to conservative treatment for low backpain. Journal of Consulting and Clinical Psychology, 35,278-284.McCulloch, J. A. (1977). Chemonucleoysis. Journal of Bone and Joint Surgery, 59, 45-52.McGill, C. M. (1968). Industrial back problems, a controlprogram. Journal of Occupational Medicine, 10, 174-78.Mahon, M. (1991). Pain perception in chronic pain patients: Asignal detection analysis. Unpublished doctoraldissertation, University of British Columbia, Vancouver,B.C.Main, C. J. & Waddell, G. (1982). Chronic pain, distress, andillness behaviour. In C.J. Main & W. R. Lindsay (Eds.),Clinical Psychology and Medicine. New York: Plenum Press.Marascuilo, L. A. (1966). Large-sample multiple comparisons.Psychological Bulletin, 65, 280-290.156Margoles, M. S. (1990). Clinical assessment and interpretationof abnormal illness behaviour in low back pain, Pain, 42,258-259.Mayer, T. G., Gatchel, R. J., Kishino, N., Keeley, J., Mayer, H.,Capra, P. & Mooney, V. (1986). A prospective short-termstudy of chronic low back pain patients utilizing novelobjective functional measurement. Pain, 25, 53-68.Melzack, R. (1975). The McGill Pain Questionnaire: Majorproperties and scoring methods. Pain, 1, 277-299.Melzack, R. & Wall, P. D. (1988). The challenge of pain. NewYork: Penguin Books.Nachemson, A. L. (1982). The natural course of low back pain.In White, A. A. & Gordon, S. L. (Eds.), Svmptoms of ideopathic low back pain (pp. 46-51). St. Louis, Mo:Mosley.Nagi, S. Z., Riley, L. E. & Newby, L. G. (1973). A socialepidemiology of low back pain in a general population.Journal of chronic disease, 26, 769-779.Nelson, M. A., Allen, P., Clamp, S. E., & De Dombal, F. T.(1979). Reliability and reproducibility of clinicalfindings in low-back pain. Spine, 4 (2), 97-101.Oostdam, E. M. M. & Duivenvoorden, H. J. & Pondaag, W. (1981).Predictive value of some psychological tests on the outcomeof surgical intervention in low back pain^patients.Journal of Psychosomatic Research, 25, 227-235.Patrick, C., Craig, K. D. & Prkachin, K. (1986). Observerjudgements of acute pain: Facial action determinants.Journal of Personality and Social Psychology, 50, 1291-1298.Philips, H. C. & Grant, L (1991). Acute back pain: Apsychological analysis. Behaviour Research and Therapy, 29,429-434.Philips, H. C. & Grant, L (1991). The evolution of chronic backpain problems: A longitudinal study. Behaviour Research andTherapy, 29, 435-441.Philips, H. C., Grant, L & Berkowitz, J. (1991). The preventionof chronic pain and disability: A preliminary investigation.Behaviour Research and Therapy, 29, 442-450.Prkachin, K. M. (1992). A cross-modal comparison of painindices. II. The consistency of facial expressions. Unpublished manuscript.157Prkachin, K. M. & Craig, K. D. (1985). Influencing non-verbalexpressions of pain: Signal detection analyses. Pain, 21,399-409.Prkachin, K. M., Currie, N. A. & Craig, K. D. (1983). Judgingnonverbal expressions of pain. Canadian Journal of Behavioural Science, 15, 43-73.Prkachin, K. M. & Mercer, S. (1989). Pain expression in patientswith shoulder pathology: validity, properties andrelationship to sickness impact. Pain, 39, 257-265.Ransford, A. 0., Cairns, D. & Mooney, V. (1976). The paindrawing as an aid to the psychologic evaluation of patientswith low back pain. Spine, 1 (2), 127-134.Redd, W. H. (1980). Stimulus control and the extinction ofpsychosomatic symptoms. Journal of Consulting and Clinical Psychology, 48, 448-455.Reesor, K. A. & Craig, K. D. (1988). Medically incongruentchronic back pain: Physical limitation, suffering, andineffective coping. Pain, 32, 35-45.Reuler, J. B., Girard, d. E., & Nardone, D. A. (1980). Thechronic pain syndrome: Misconceptions and management.Annals of Internal Medicine, 93, 588-596.Rosen, J. C., Frymoyer, J. W. & Clements, J. H. (1980). Afurther look at validity of the MMPI with low back patients.Journal of Clinical Psychology, 36, 994-1000.Rosenteil, A. & Keefe, F. (1983). The use of coping strategiesin chronic low back pain patients: Relationship to patientcharacteristics and adjustment. Pain, 17, 33-43.Rosomoff, H. L., Fishbain, D. A., Goldberg, M., Santana, R. &Rosomoff, R. S. (1989). Physical findings in patientswith chronic intractable benign pain of the neck and/ orback. Pain, 37, 279-287.Rudy, T. E., Turk, D. C. Brena, S. F. Stieg, R. L. & Brody, M. C.(1990). Quantification of biomedical findings of chronicpain patients development of an index of pathology. Pain,42, 167-182.Sternbach, R. A. (1974). Pain patients, traits, and treatments.New York: Academic Press.Sternbach, R. A. (1985). Acute versus chronic pain. In P. D.Wall & R. Melzack (Eds.), Textbook of pain (pp. 173-177).Edinburgh: Churchill Livingstone.158Sternbach, R. A. & Tursky, B. (1965). Ethnic differences^amonghousewives' psychological and skin potential responses toelectric shock. Psychophysiology, 1, 241-246.Sternbach, R. A., Wolf, S. R., Murphy, R. W. & Akeson, W.^H.(1973). Traits of pain patients: The low-back 'loser'.Psychosomatics, 14, 226-229.Swalm, D. & Craig, K. D. (1990). Differential impact of placeboon verbal and nonverbal measures of pain in men and women.Unpublished manuscript.Swanson, D. W. & Maruta, T., (1980). Patients complaining ofextreme pain. Mayo Clinic Proceedings, 55, 563-566.Taylor, W. P., Stern, W. R. & Kubiszyn, T. W. (1984). Predictingpatients perceptions of response to treatment for low-backpain. Spine, 9, 313-316.Teske, K., Daut, R. L. & Cleeland, C. S. (1983). Relationshipbetween nurses' observations and patients' self report ofpain. Pain, 16, 289-296.Trief, P. M., Elliott, D. J. & Frederickson, B. E. (1987).Functional vs. organic pain: A meaningful distinction?Journal of Clinical Psychology, 43, 219-226.Turk, D. C., Meichenbaum, D. & Genest, M. (1983). Pain andbehavioural medicine: A cognitive-behavioural perspective.New York: Guilford Press.Turk, D. C. & Rudy, T. E. (1985). Pain experience: Assessing thecognitive component. Paper presented at the annual meetingof American Pain Society, Dallas, TX.Turk, D. C. & Rudy, T. E. (1987). Towards a comprehensiveassessment of chronic pain patients. Behaviour Research andTherapy, 25 (4), 237-249.Turner, J. A. & Clancey, S. (1986). Strategies for coping withchronic low back pain: Relationship to pain and disability.Pain, 24, 355-404.Uden, A., Astrom, M. & Bergenudd, H. (1988). Pain drawings inchronic back pain. Spine, 13 (4), 398-392.von Baeyer, C. L. (1982). Repression-sensitization, stress, andperception of pain in others. Perceptual and Motor Skills,55, 315-320.von Baeyer, C. L., Bergstrom, K. J., Brodwin, M. G. & Brodwin, S.K. (1983). Invalid use of Pain Drawings in psychologicalscreening of back pain patients. Pain, 16, 103-107.159von Baeyer, C. L., Johnson, M. & McMillan, M. (1982).Consequences of nonverbal expression of pain: Patientdistress and observer concern. Social Science and Medicine,19 (12), 1319-1324.Waddell, G., (1982). An approach to backache. British Journal of Hospital Medicine, 22, 187-219.Waddell, G., (1987). Clinical assessment of lumbar impairment.Clinical Orthopaedics and Related Research, 221, 110-120.Waddell, G., Bircher, M., Finlayson, D. & Main C. J., (1984).Symptoms and signs: Physical disease or illness behaviour?British Medical Journal, 289, 739-741.Waddell, G. & Main, C. J. (1984). Assessment of severity in low-back pain disorders. Spine, 9 (2), 204-208.Waddell, G., Main, C. J., Morris, E. W., Di Paolo, M. & Gray, I.(1984). Chronic low-back, psychologic distress, and illnessbehaviour. Spine, 9 (2), 209-213.Waddell, G., McCulloch, J. A., Kummel, E. & Venner, R. M. (1980).Nonorganic physical signs in low-back pain. Spine, 5, (2),117-125.Waddell, G., Pilowsky, I. & Bond, M. R. (1989). Clinical^assessment and interpretation of abnormal illness behaviour inlow back pain. Pain, 39, 41-53.Waddell, G., Pilowsky, I. & Bond, M. R. (1990). Non-anatomicalsymptoms and signs: A response to critics. Pain, 42, 260-261.Wall, P. D. (1979). On the relation of injury to pain. The JohnJ. Bonica lecture. Pain, 6, 253-264.Wolff, B. B. (1978). Behavioural measurement of human pain. InR. A. Sternbach, (Ed.), The psychology of pain. New York:Raven Press.160APPENDIX ANonorganic Physical Signs(Waddell et al., 1980)A) Overreaction to examinationFacial expressionMuscle tensionCollapsingSweatingB) TendernessSuperficialNonanatomicC) SimulationAxial loadingSimulated rotationD) DistractionStraight leg raisingE) Regional DisturbancesWeaknessSensoryTotal ScoreScoringAny individual sign counts as a positive sign for that type;a finding of two or more of the five types is clinicallysignificant. Isolated positive signs are ignored.APPENDIX BPain Drawing(Ransford et al., 1976)On the human form below, mark where your numbness or pain is,using the kind of marks that correspond to what you feel in eacharea.Numbness^Pins & Needles 0000^Aching ^^^^Burning xxxx^Stabbing ////161Scoring: Pain Drawing (From Ransford et al., 1976)Points are assigned on the following basis:1. Unreal drawings:Poor anatomical localization, scores 2 unless indicated;bilateral pain not weighted unless indicated.a) total leg painb) lateral whole leg pain (trochanter area and lateral thighallowed)c) circumferential thigh paind) bilateral anterior tibial area pain (unilateral allowed)e) circumferential foot pain (scores 1)f) bilateral foot pain (scores 1)g) use of four or five modalities suggested in instruction(patient is unlikely to have burning areas, stabbingareas, pins and needles, numbness and aching alltogether; scores 1)2. Drawings showing "expansion" or magnification" of pain:This may also represent unrelated symptomatology; bilateral painnot weighted.a) back pain radiating to illiac crest, groin, or anteriorperineum (each scores 1; coccygeal pain allowed)b) anterior knee pain (scores 1)c) anterior ankle pain (scores 1)d) pain drawing outside the outline; this is a particularlygood indication of magnification (scores 1 or 2 dependingon extent)3. "Particularly Hurt Here" indicators:Each category scores 1; multiple use of a category is notweighted.a) add explanatory notesb) circle painful areasc) draw lines to demarcate painful areasd) use arrowse) go to excessive trouble and detail in demonstrating thepain areas using the symbols suggested4. "Look How Bad I Am" indicators:Additional painful areas in the trunk, head, neck or upperextremities drawn in. Tendency toward total body pain (scores 1if limited to small areas, otherwise 2).162APPENDIX CInappropriate Symptom Inventory(Waddell et al., 1984)Note: Must rule out spinal pathology (e.g., tumor or infectionsin spinal column) before the symptoms listed below are consideredto be inappropriate.Unless otherwise indicated score 1 if the patient responds yes tothe questions.Interview Questions:1). Do you get pain at the tip of your tailbone ? ^2). Does your whole leg ever become painful ?3). Does your whole leg ever go numb ?4). Does your whole leg ever give way ?5). In the past year have you had any spells withvery little pain ? (no is scored as 1)6). Do you think that the treatments you have hadso far for your pain problem have helped you ?(no is scored as 1)7). Have you ever had to go to the emergencydepartment in a hospital for your backpain ?Total163164APPENDIX DConsent Form: Nonverbal Expression StudyWe are interested in studying how patients respond to thevarious assessment procedures that are carried out byphysiotherapists. For our study we are requesting yourpermission to videotape a short segment of the physiotherapysession you are about to undergo. During this segment of thephysiotherapy session we will be asking you to carry out certainmovements under different instructions. In order to assess yourreactions to the procedures, the videotapes will be coded andviewed by independent observers. You will also be requested tofill out several brief questionnaires related to the discomfortyou may experience during those movements and more generally on aday-to-day basis.We are also interested in understanding how back painaffects your life, and how you have come to cope with thediscomfort you may be feeling. Therefore, we will also be askingyou to fill out a number of questionnaires concerning your painexperience and your coping strategies.The short segment of the physiotherapy session that I aminterested will take about five minutes. The questionnaires canbe completed during your treatment session, and will take about15 minutes. All of the information will be confidential, and youwill be identified only by number to ensure your anonymity. Wewould greatly appreciate your cooperation; however, you are freeto refuse and doing so will in no way affect your treatment atthe Lansdowne Physiotherapy Clinic. Thank you.Heather Hadjistavropoulos, M.A. StudentDepartment of PsychologyUBC Ph. 822-5280Kenneth D. Craig, Ph.DDepartment of PsychologyUBC Ph. 822-3948I agree to participate in this study and give permission to theLansdowne Physiotherapy Clinic to release medical informationsolely for the purposes of this investigation, and subject to thecondition that this information is kept in strict confidence. Iam aware that I can stop participation at any time withoutpenalty.Signature ^Date165APPENDIX EPhysiotherapy ProtocolNEUTRAL: I would like you to keep a neutral expression on yourface for about 10 seconds.SAY OKAY AFTER 10 SECONDSBASELINE: I would like you to wiggle your toes for about 10seconds.SAY OKAY AFTER 10 SECONDSTHE NEXT THREE SETS OF INSTRUCTIONS ARE TO BE PRESENTEDIN RANDOM ORDERGENUINE: I would like you to lift both of your legs up to abouthere (10-12 inches off of the examination table), and hold it forabout 10 seconds. When you are doing this please genuinelyexpress the pain or discomfort you are feeling.SAY OKAY AFTER 10 SECONDSMASKED: I would like you to lift both of your legs to about here(10-12 inches off of the examination table), and hold it forabout 10 seconds. When you are doing this try and mask or coverup any pain you may be feeling.SAY OKAY AFTER 10 SECONDSEXAGGERATED: I would like you to lift both of your legs to abouthere (10-12 inches off of the examination table), and hold it forabout 10 seconds. When you are doing this exaggerate thediscomfort or pain you are feeling.SAY OKAY AFTER 10 SECONDS166APPENDIX FDebriefing ScriptThank you for your participation in this study. The purposeof this study is twofold. First, we are interested in examiningnonverbal expressions of pain in response to the procedurescarried out by physiotherapists. As you may well know, whenassessing pain in others, health care professionals often mustrely heavily on nonverbal information. By nonverbal informationwe mean expressions on your face. It is important that we studynonverbal expressions in order to ensure that an accurateassessment of the discomfort and pain one feels will be made.Also of interest in this study is how ongoing paininfluences how you feel and how you learn to cope with the pain.Since in many instances it is difficult to treat back pain,learning to cope with pain appears to be one of the best possiblesolutions for dealing with pain. We are hoping that theinformation that we obtained from you will help us learn moreabout the various ways of dealing with acute and chronic pain,and will also help us determine if there is a relationshipbetween the way one copes with pain and the way one expressespain nonverbally.If you are interested in further information we would behappy to answer your questions. Once again thank you for yourcooperation.Heather Hadjistavropoulos, M.A. StudentDepartment of PsychologyUBC Ph. 822-5280Kenneth D. Craig, Ph.DDepartment of PsychologyUBC Ph. 822-3948APPENDIX GClinical Chart for Routine Calculation of Impairment(Waddell,^G.,^& Main,^C.^J.,^1984)Mathematical Constant 28Pain Problem Back pain 0Back and referred pain 8Root pain -2Neurogenic claudication 8Time pattern Recurring 4Chronic 8Previous fracture Transverse process 1Compression 2Fracture/dislocation 6Previous back surgery None 0One 3More than one 6Root Compression None 0Doubtful 1Definite 2Subtotal A +Lumbar flexion cm x 2Straight leg raising ^/10(left)Straight leg raising ^/10(right)Subtotal B167+Subtotal ASubtotal BImpairmentAPPENDIX HThe Oswestry Low Back Pain Disability Questionnaire (OLDQ)(Fairbank, Couper, Davies, & O'Brien, 1980)Please read:This questionnaire has been designed to provide information as tohow your back pain has affected your ability to manage ineveryday life. Please answer every section, and mark in eachsection only the one statement which applies to you. We realizeyou may consider that two statements in any one section relate toyou, but please just mark the statement which most clearlydescribes your problem.Pain Intensity I can tolerate the pain I have without using pain killers.The pain is bad but I manage without taking pain killers.Pain killers give complete relief from pain.Pain killers give moderate relief from pain.Pain killers give very little relief from pain.Pain killers have no effect on the pain; I do not use them.Personal Care (Washing, Dressing, etc.) I can look after myself normally without causing extra pain.I can look after myself but it causes extra pain.It is painful to look after myself; I am slow and careful.I need some help but manage most of my personal care.I need help every day in most aspects of self care.I do not get dressed, wash with difficulty and stay in bed.LiftingI can lift heavy weights without causing extra pain.I can lift heavy weights but it gives extra pain.Pain prevents me from lifting heavy weights off the floor, but Ican manage if they are conveniently positioned.Pain prevents me from lifting heavy weights but I can managelight to medium weights if they are conveniently positioned.I can lift only very light weights.I cannot lift or carry anything at all.WalkingPain does not prevent me from walking any distance.Pain prevents me from walking more than 1 mile.Pain prevents me from walking more than 1/2 mile.Pain prevents me from walking more than 1/4 mile.I can only walk using a stick or crutches.I am in bed most of the time, and have to crawl to the toilet.168Sitting I can sit in any chair as long as I like.I can only sit in my favorite chair as long as I like.Pain prevents me from sitting more than 1 hour.Pain prevents me from sitting more than 1/2 hour.Pain prevents me from sitting more than 10 minutes.Pain prevents me from sitting at all.StandingI can stand as long as I want without extra pain.I can stand as long as I want but it gives extra pain.Pain prevents me from standing for more than 1 hour.Pain prevents me from standing for more than 30 minutes.Pain prevents me from standing for more than 10 minutes.Pain prevents me from standing at all.SleepingPain does not prevent me from sleeping well.I can sleep well only by using tablets.Even when I take tablets I have less than six hours sleep.Even when I take tablets I have less than four hours sleep.Even when I take tablets I have less than two hours sleep.Pain prevents me from sleeping at all.Sex LifeMy sex life is normal and causes no extra pain.My sex life is normal but increases the degree of pain.My sex life is nearly normal but is very painful.My sex life is severely restricted by pain.My sex life is nearly absent because of pain.Pain prevents any sex life at all.Social LifeMy social life is normal and gives me no extra pain.My social life is normal but increases the degree of pain.Pain has no significant effect on my social life apart fromlimiting more energetic interests (e.g., dancing, etc).Pain has restricted my social life and I do not go out as often.Pain has restricted my social life to my home.I have no social life because of pain.TravellingI can travel anywhere without extra pain.I can travel anywhere but it gives me extra pain.Pain is bad but I manage journeys over two hours.Pain restricts me to journeys of less than one hour.Pain restricts me to short necessary journeys under 30 minutes.Pain prevents me from travelling except to the doctor orhospital.169Scoring of the OLDO(Fairbank, Couper, Davies & O'Brien, 1980)Each section is scored on a 0-5 scale, 5 representing thegreatest disability. The scores for all sections are addedtogether, giving a possible score of 50. The total is doubledand expressed as a percentage. If a patient marks twostatements, the highest scoring statement is recorded as a trueindication of his disability. If a section is not completebecause it is inapplicable (e.g., sex life), the final score isadjusted to obtain a percentage.170APPENDIX IDescriptor Differential Scales(Gracely, 1980)Please circle one word or word-pair on each of the scales belowto describe:1) the pain you experience on an average day.2) the most severe pain you experienced in response to themovements.Sensory^ Unpleasantness A. Extremely Intense^A. Very IntolerableB. Very Intense^ B. IntolerableC. Intense^ C. Very DistressingD. Strong D. Slightly IntolerableE. Slightly Intense^E. Very AnnoyingF. Barely Strong F. DistressingG. Moderate^ G. Very UnpleasantH. Mild H. Slightly DistressingI. Very Mild^ I. AnnoyingJ. Weak^ J. UnpleasantK. Very Weak K. Slightly AnnoyingL. Faint^ L. Slightly UnpleasantM. No Sensation of Pain^M. No Discomfort171Scoring of(Gracely,Sensorythe DescriptorDubnerDifferential Scales& McGrath,^1979)UnpleasantnessA. Extremely Intense 59.5 A. Very Intolerable 44.8B. Very Intense 43.5 B. Intolerable 32.8C. Intense 34.6 C. Very Distressing 18.3D. Strong 22.9 D. Slightly Intolerable 13.6E. Slightly Intense 21.3 E. Very Annoying 12.1F. Barely Strong 12.6 F. Distressing 11.4G. Moderate 12.4 G. Very Unpleasant 10.7H. Mild 05.5 H. Slightly Distressing 06.2I. Very Mild 03.9 I. Annoying 05.7J. Weak 02.8 J. Unpleasant 05.6K. Very Weak 02.3 K. Slightly Annoying 03.5L. Faint 01.1 L. Slightly Unpleasant 02.8M. No Sensation 00.0 M. No Discomfort 00.0172173APPENDIX JCoping Strategy QuestionnaireIndividuals who experience pain have developed a number of waysto cope, or deal with, their pain. These include saying thingsto themselves when they experience pain, or engaging in differentactivities. Below are a list of things that patients havereported doing when they feel pain. For each activity, pleaseindicate, using the scale below, how much you engage in thatactivity when you feel pain, where a 0 indicates you never dothat when you experience pain, a 3 indicates that you sometimesdo that when you experience pain, and a 6 indicates that youalways do that when you experience pain. Remember you can useany point along the scale.0^1^2^3^4^5^6Never Sometimes Alwaysdo that^do that^do thatWHEN I FEEL PAIN ....1. I try to feel distant from the pain, almost as if the painwas in somebody else's body.2. I leave the house and do something, such as going to themovies or shopping.3. I try to think of something pleasant.4. I don't think of it as pain but rather as a dull or warmfeeling.5. It is terrible and I feel it is never going to get anybetter.6. I tell myself to be brave and carry on despite the pain.7. I read.8. I tell myself that I can overcome the pain.9. I count numbers in my head or run a song through my mind.10. I just think of it as some other sensation such as numbness.11. It is awful and I feel it overwhelms me.12. I play mental games with myself to keep my mind off the pain.13. I feel my life isn't worth living.14. I know someday someone will be here to help me and it will goaway.15. I pray to God it won't last long.16. I try not to think of it as my body, but rather as somethingseparate from me.17. I don't think about the pain.18. I try to think years ahead, what everything will be likeafter I've gotten rid of the pain.19. I tell myself it doesn't hurt.20. I tell myself I can't let the pain stand in the way of what Ihave to do.21. I don't pay any attention to it.22. I have faith in doctors that someday there will be a cure formy pain.17423. No matter how bad it gets I know I can handle it.24. I pretend it is not there.25. I worry all the time about whether it will end.26. I replay in my mind pleasant experiences in the past.27. I think of the people I enjoy doing things with.28. I pray for the pain to stop.29. I imagine that the pain is outside of my body.30. I just go on as if nothing has happened.31. I see it as a challenge and don't let it bother me.32. Although it hurts I just keep going.33. I feel I can't stand it any more.34. I try to be around other people35. I ignore it.36. I rely on my faith in God.37. I feel like I can't go on.38. I think of things I enjoy doing.39. I do anything to get my mind off the pain.40. I do something I enjoy, such as watching TV or listening tomusic.41. I pretend it is not a part of me.42. I do something active, like household chores or projects.Based on all of the things you do to cope, or deal with, yourpain, on an average day, how much control do you feel you haveover it? Please circle the appropriate number. Remember, you cancircle any number along the scale.^0^1^2^3^4^5^6No Some CompleteControl^Control^ControlBased on all of the things you do to cope, or deal with yourpain, on an average day, how much are you able to decrease it?Please circle the appropriate number. Remember, you can circleany number along this scale.0^1^2^3^4^5^6Can't decrease Can decrease^Can decreaseit at all^it somewhat^it completelyKey to Coping Strategy QuestionnaireCognitive Coping Strategies: 1. Diverting Attention: 3, 9, 12, 26, 27, 382. Reinterpretation:^1, 4, 10, 16, 29, 413. Catastrophizing: 5, 11, 13, 25, 33, 374. Ignoring Sensations: 17, 19, 21, 24, 30, 355. Praying or Hoping: 14, 15, 18, 22, 28, 366. Coping Self-Statements:^6, 8, 20, 23, 31, 32Behavioural Coping Strategies:1. Increased Behavioural Activities: 2, 7, 34, 39, 40, 42Effectiveness Ratings:1. Control Over Pain2. Ability to Decrease Pain175176APPENDIX KPain Eexperience ScaleMany people report having the following kinds of thoughts andfeelings when their pain is very severe. We would like to knowhow frequently you experience each of the thoughts and feelingslisted below when your pain is very severe. Read each item andthen circle a number on the scale under the statement to indicatehow often you have that thought or feeling.1. I feel frustrated.0^1^2^3^4^5^6Never Very Often2. I think about my pain getting worse.0^1^2^3^4^5^6Never Very Often3. I feel irritable.0^1^2^3^4^5^6Never Very Often4. I am depressed because of my pain.O 1^2^3^4^5^6Never Very Often5. I wonder what it would be like to never have any pain.O 1^2^3^4^5^6Never Very Often6. I feel angry.O 1^2^3^4^5^6Never Very Often7. I feel overwhelmed.O 1^2^3^4^5^6Never Very Often8. I feel afraid that my pain will get worse.O 1^2^3^4^5^6Never Very Often9. I think, " This pain is driving my crazy"0^1^2^3^4^5^6^Never Very Often10. I feel impatient with everybody.0^1^2^3^4^5^6Never Very Often11. I worry about my family.0^1^2^3^4^5^6Never Very Often12. I think about whether life is worth living.0^1^2^3^4^5^6Never Very Often13. I feel anxious.0^1^2^3^4^5^6Never Very Often14. I feel disappointed with myself for giving into the pain.O 1^2^3^4^5^6Never Very Often15. I feel everyone is getting on my nerves.O 1^2^3^4^5^6Never Very Often16. I think, "It is so hard to do anything when I have pain."O 1^2^3^4^5^6Never^ Very Often17. I wonder how long this will last.0^1^2^3^4^5^6Never Very Often18. I think of nothing other than my pain.0^1^2^3^4^5^6Never Very Often17719. I feel sorry for myself.0^1^2^3^4^5^6Never Very Often178Scoring of the Pain Experience ScaleScale 1: Emotionality = (1, 3, 4, 6, 7, 9, 10, 12, 13, 15, 16,18, 19) /13Scale 2: Worry = (2, 5, 8, 11, 14, 17) /6179APPENDIX LScoring Units for the Facial Action Coding System(Ekman & Friesen,^1978b)Upper Face^ Lower FaceAU Description AU Description1 Inner Brow Raise 9/10 Levator Contraction2 Outer Brow Raise * 11 Nasolabial Deepen4 Brow Lower 12 Lip Corner Pull5 Upper Lid Raise 13 Cheek Puff6/7 Orbit Tightening 14 Dimpler41 Lids Droop 15 Lip Corner Depress42 Eye Slit 16 Lower Lip Depress43 Eyes Closed 17 Chin Raise44 Squint 18 Lip Pucker45 Blink 20 Lip Stretch46 Wink 22 Lip Funnel23 Lips Tight24 Lip Press25/26/27 Mouth Opening28^Lip SuckMiscellaneous ActionsAU Description AU Description8 Lips Toward 33 Blow19 Tongue Show 34 Puff21 Neck Tighten 35 Cheek Suck29 Jaw Thrust 36 Tongue Bulge30 Jaw to Side 37 Lip Wipe31 Jaw Clench * 38 Nostril Dilate32 Bite 39 Nostril CompressNote. Those AUs marked with an asterisk were not coded forintensity.180181Baseline^ MaskedGenuine ExaggeratedFigure :1. Artist's drawing of one patient's facial activityduring baseline, masked, genuine, and exaggerated painsegments.