Detecting sincerity of effort: a summary of methods and approaches.

Key words: Malingering; Neck and trunk, back; Pain; Reproducibility; Sincerity of effort; Submaximal effort.

Sincerity of effort, in this article, refers to a patient's conscious motivation to perform optimally during an evaluation.[1-3] A sincere effort is the patient's best or optimal physical performance, whereas an insincere effort is one in which the patient deliberately gives less than a full effort during physical examinations.[1-3] Patients whose efforts are not sincere during evaluation may overuse treatment, may have a prolonged recovery, may have an increased cost of care, or may receive unwarranted disability payments.[1-3] For obvious reasons, there is a keen interest in sincerity of effort from the medicolegal community and from the insurance industry.[1,3,4] Sincerity of effort during evaluation of patients with low back pain (LBP) is frequently questioned when the patient's injury occurs in a work-related environment or when injury-related litigation is pending. Employers, third-party payers, attorneys, and case managers want to know whether the patient with LBP is giving a maximal or best effort during impairment and functional evaluation.

In response to the demand for information on patient participation or cooperation during evaluations, several methods are now widely used by clinicians to evaluate sincerity of effort (Tab. 1).[1-21] These methods are often interpreted by referral sources as measures of sincerity of effort, even when explicit detection of sincerity of effort is not stated in therapists' or clinicians' reports. Do we have evidence in support of the reliability and validity of measurements obtained with these methods for the purpose of detecting sincerity of effort? Are we relying, instead, on traditional clinical opinion when we use these tests? For what purpose were these tests developed? If the tests give unreliable or inaccurate measurements, patients with lumbar dysfunction can be labeled inappropriately and negatively. Such labeling results in misdiagnosis, improper treatment, increased litigation, and increased cost of care. Our credibility as a profession rests on our ability to select and use reliable and valid clinical measures.

Table 1. Widely Used Methods of Determining Sincerity of Effort

Waddell's nonorganic signs[29] Coefficient of variation[84] Bell-shaped curve[113] Rapid exchange grip[106] Correlation between musculoskeletal evaluation and functional

capacity evaluation[17] Documentation of pain behavior[54] Documentation of symptom magnification[9] Ratio of heart rate and pain intensity[119]

We argue that our credibility is enhanced when we use appropriate terminology to address the issue of sincerity of effort and to differentiate it from the biobehavioral aspects of delayed recovery. The term "validity," for example, often is used to address sincerity of effort.[6,7,10,11,14] Test results are described as "valid" or "invalid" based on the results of a series of tests that are alleged to test sincerity of effort. Use of the term "validity" to describe sincerity of effort is an inappropriate application of a scientific term.[22] Validity refers to the extent to which a measure can be used to make an inference or judgment,[23] Research establishes evidence in support of validity, which does not change with the patient's level of effort. If patients do not give a full effort during physical evaluations, then the test results represent what they were willing to do. Because patients cannot be forced to expend more effort than they are willing to expend, validity is unaffected. There is no evidence reported in the peer-reviewed literature that any of the tests designed to provide a "validity" profile of the patient are valid in the scientific sense.[22]

Other terms that may cloud our understanding of the biobehavioral factors affecting recovery include "symptom magnification" and "exaggerated pain behavior." These terms are frequently used in clinical practice, in standardized evaluations, and in the literature to identify patients who are thought to be exaggerating the severity of their medical conditions.[9,11,13-15,19,24] The use of this terminology is not theoretically sound. By definition, a symptom is a "sensation experienced by the patient."[25] There is no laboratory test or imaging technique that can measure the patient's true versus reported experience of sensation. The terms "magnification" and "exaggeration" imply that we can measure true sensations and compare these measurements with patient reports. Thus, by definition, "symptom magnification" and "exaggerated pain behavior" cannot be measured. Use of these terms, therefore, should be avoided, as they add little information that leads to improved treatment for the patient with delayed recovery.[26]

Some patients with musculoskeletal dysfunction can give less than a full effort during physical evaluations for a variety of reasons. Pain, fear of pain, fear of reinjury, anxiety, depression, lack of understanding of instructions, lack of understanding of the importance of the test, and secondary financial gain are some of the reasons underlying self-limiting behavior. Lechner et al[27] found that therapists can reliably identify when a patient apparently is giving a full effort during functional capacity evaluations (FCEs). These investigators identified maximal effort by comparing patients' willingness to continue performing functional tasks with therapists' observations of body alignment and movement patterns. Patients were classified as participating fully, stopping before a full effort was reached, or willing to continue beyond a safe maximum effort. Sixty-two percent of the subjects self-limited their performance on 2 or more of the tasks tested, in our view, giving clinicians an adequate opportunity to judge and score self-limiting behavior. Using this approach, they achieved kappa scores for interrater reliability ranging from .56 to .97, with 95% of the scores being greater than .61.[27] According to Landis and Koch's suggested reference values, these kappa scores fall in the substantial to almost perfect agreement range.[28]

Lechner et al,[27] however, made no claim that underlying motivation was measured. If self-limiting behavior predominates during an FCE, further psychosocial or environmental evaluation may elucidate underlying factors that can be addressed during treatment, thus optimizing the likelihood of patient recovery.[26] Lack of sincerity of effort, however, should not be inferred from self-limiting participation in testing.

Waddell's Nonorganic Signs

Waddell et al,[29] in 1980, first described nonorganic signs as clinical signs that have a "predominantly nonorganic basis" in patients with LBP. They defined 8 tests for these signs and grouped them into 5 types (Tab. 2). In types with more than one test, scoring positively on one test yields a positive score for that type. According to Waddell et al,[29] a patient must score positively on 3 out of the 5 types of signs to score positively on the nonorganic signs. Waddell et al[29] stated that scoring positively on these signs identifies patients who might benefit from psychological assessment.

Table 2.
Waddell's Nonorganic Signs (NOS) Test Administration and Scoring(a)

Type of NOS             Sign

Tenderness              Superficial

                        Deep

Simulation tests        Axial loading

                        Rotation

Distraction test        Straight leg raise
                         (flip test)

Regional disturbances   Weakness

                        Sensory

Overreaction            Overreaction

Type of NOS             Test Administration

Tenderness              Examiner pinches skin lightly
                         between thumb and forefinger
                         over the thoraco-lumbo-sacral
                         region

                        Examiner applies deep pressure over
                         the thoraco-lumbo-sacral region
Simulation tests        With patient standing, the examiner
                         applies pressure to the top of the
                         patient's head

                        With patient standing, the examiner
                         rotates the shoulders and pelvis as
                         a unit

Distraction test        Straight leg raise is tested in a
                         supine position, and the patient is
                         then asked to extend the knee in a
                         sitting position

Regional disturbances   Manual muscle testing of the lower
                         extremities

                        Sensory testing for light touch, and
                         pinprick over the lower extremities

Overreaction            Examiner makes observations during
                         the examination

Type of NOS             Positive Score

Tenderness              Patient reports tenderness over a "wide
                         area of lumbar skin"

                        Patient reports tenderness over a "wide
                         area"
Simulation tests        Patient reports back pain secondary to
                         the pressure

                        Patient reports back pain secondary to
                         the rotation

Distraction test        Patient "shows marked improvement"
                         when sitting knee extension is
                         compared with supine
                         straight-leg-raising test
Regional disturbances   "Giving way" of many muscle groups
                         that cannot be explained on a
                         localized neurological basis
                        "Diminished sensation to light touch,
                         pinprick... fitting a `stocking' rather
                         than dermatomal pattern"
Overreaction            Disproportionate verbalization, facial
                         expression, muscle tension and
                         tremor, collapsing, or sweating

These signs were not intended for use in detecting sincerity of effort or malingering.[29,30] Waddell et al[29] reported that the nonorganic signs did not correlate with the F and K validity scores of the Minnesota Multiphasic Personality Inventory in 120 patients with chronic LBP from Canada and Great Britain. These investigators stated that the nonorganic signs were "not limited to, nor specific to, medicolegal and compensation situations."[29] Unfortunately, nonorganic signs are frequently used in clinical practice, in standardized evaluations, and in regulatory guidelines to imply sincerity of effort or an exaggeration of symptom reporting.[7,11,13-15] In addition, clinical investigators have strayed from the original intent of the nonorganic signs and have described their purpose as being to "detect whether patients were accurately reporting pain"[31] or to "establish the authenticity or validity of pain reports."[32]

Even when nonorganic signs are used appropriately, the premise that they are based on nonorganic symptoms is questionable. For example, the tenderness tests are scored positively if the superficial or deep tenderness does not follow a pattern associated with nerve root irritation. Regional disturbances are scored positively if the findings do not follow a dermatomal distribution.[29] Doxey et al[33] demonstrated that the nonorganic signs were associated with the absence of nerve root involvement. Such an approach seems to imply that the only valid organic source of pain is the nerve root. There is, however, considerable evidence that many structures other than the nerve root can lead to pain that extends over a broad area.[34-37] Thus, the tenderness tests may contribute to an erroneous classification of patients as needing further psychological assessment when their primary problem stems from organic sources or tissues other than nerve roots.

Several issues surrounding test administration should cause us to question the validity of nonorganic signs for detecting nonorganic problems. The simulation tests are assumed not to cause pain in the lumbar region.[29] In our clinical experience, patients with poor trunk strength or a lack of kinesthetic awareness may respond to axial loading with movement and resulting pain. The trunk rotation test presumes that the examiner is able to produce simultaneous rotation at the pelvis and shoulders without rotation occurring at the lumbar spine. In our clinical experience, eliminating motion in the lumbar region during this maneuver may not be possible, even with precise verbal instructions and careful hand placement.

The supine straight-leg-raising test and knee extension in a sitting position are presumed by Waddell et al[29] and by many clinicians to be equivalent measures. In fact, the measures are quite different. In a supine position, the spine and pelvic positions are supported by the surface on which the patient is lying, providing some stability for the lumbar spine and pelvis. In sitting, however, the spine and pelvis are unsupported.[29] The patient may slump or allow the pelvis to rotate posteriorly when the knee is extended, particularly if hamstring muscle tightness is present. Thus, in our clinical experience, knee extension in a sitting position may exceed the range of motion during the supine straight-leg-raising test in some patients who are giving a sincere effort during musculoskeletal evaluation. Although many clinicians consider only "profound" differences to be positive, the cutoff point has not been defined or examined in research studies.[29]

According to our interpretation, Waddell's overreaction test, another of the nonorganic signs, is based on the assumption that there is a standard, acceptable intensity of response to the experience of pain, against which reaction can be measured.[29] We know that past experiences, cultural background, and socioeconomic status can affect the intensity of the pain experience and expressions of pain.[38-42] Both the examiner and the patient are subject to these influences. Perhaps these factors were, at least in part, responsible for the poor reliability of this nonorganic sign.[43]

Several aspects of the nonorganic signs test administration and interpretation appear to lack clear definitions. Terms such as "wide area" in the assessment of tenderness and "disproportionate" in the overreaction test are 2 examples of terms lacking operational definitions that, in theory, could enhance the reliability and consistency of scoring the responses. Techniques of test administration, such as the amount of pressure exerted during the axial loading, hand placement during rotation, and pelvic control during the distraction test, are not standardized in Waddell's original description of the nonorganic signs[29] and may introduce variability into testing and scoring the nonorganic signs. These clinical opinions of the nonorganic signs will require further study.

Five studies addressing the reliability of scores for nonorganic signs were found (Tab. 3). Early studies conducted by Waddell and colleagues[29,44] demonstrated a high degree of interrater reliability. This reliability, however, was established among individuals who received considerable informal, nonstandardized training from Waddell. The ability to generalize the results of this study to routine clinical practice, therefore, is limited. Other investigators[43,45] have had more difficulty demonstrating the reliability of scores for nonorganic signs. The study by McCombe et al[45] involved a low proportion of positive nonorganic signs, which may have artificially lowered the kappa coefficient. Regardless of these statistical issues, these authors advise caution in interpreting superficial tenderness and abnormal sensory or motor disturbance. Future studies should report both kappa values and percentages of agreement to address this issue. Spratt et al[32] improved the reliability of scores for nonorganic signs by using a system for coding the behavioral responses to 4 of the nonorganic signs tests: rotation, axial loading, superficial tenderness, and distraction. Their design, however, allowed 2 clinicians to observe a patient simultaneously. Such a design addresses only the reliability of interpreting responses, not the ability to obtain the measurement in clinical practice. Because much of the variability of the non-organic signs may lie in test administration, reliability of the behavioral classification may have been overstated in this study.

Table 3
Studies Examining Reliability of Noting the Presence
of Nonorganic Signs (NOS)

Author (Year)         No. of      Type of
                      Subjects    Subjects

Waddell                  50       Canadian WCB(a)
  et al(29) (1980)

Waddell                  50       WCB patients
  et al(44) (1982)

Korbon                   39       Outpatients with
et al(43,b) (1987)                  chronic LBP(c)

McCombe              Group1:50    Outpatients with
et al45 (1989)       Group2:33      LBP

Spratt et al(32)         42       LBP>4 mo
(1990)

Author (Year)                   Method

Waddell              Interrater reliability between 2
  et al(29) (1980)     physicians who independently
                       examined patients within same
                       week

                     Physicians "had worked closely
                       together for >6 mo"

                     Test-retest reliability between
                       admission and discharge scores
                       (average of 23 d between
                       evaluations)

Waddell              Patients examined independently by
  et al(44) (1982)     2 physicians

                     Before data collection, "detailed
                       discussion on exact information to
                       be gathered, and its format,
                       qualifications, and exclusions"

Korbon               Studied interrater reliability using 2
et al(43,b) (1987)     physicians who evaluated patients
                       independently

                     Attempted more quantitative version
                       of NOS

McCombe              Interrater reliability between 2
et al45 (1989)         orthopedic surgeons and between
                       an orthopedic surgeon and a
                       physical therapist

Spratt et al(32)     Rater pairs (one examiner, one
(1990)                 observer) evaluated rotation,
                       superficial tenderness, axial
                       loading, distraction; determined
                       reliability of judging behavioral
                       response using standardized
                       ordinal scoring system

Author (Year)           Findings

Waddell              Interrater 80% agreement
  et al(29) (1980)   Intrarater 85% agreement for
                       identifying positive versus
                       negative NOS

Waddell              Moderate to substantial
  et al(44) (1982)     reliability in assessing
                       NOS ([Kappa]=.55-.77)

                     Did not report statistics for
                       reliability for each sign

Korbon               Found axial loading (.69),
et al(43,b) (1987)     rotation (.57), and
                       overreaction (.44) to be
                       unreliable

                     Documenting the degree of
                       tenderness (.48), weakness
                       (.72), and sensory
                       disturbance (.83) also
                       unreliable

McCombe              Found tenderness ([[Kappa].sub.1] =.29,
et al45 (1989)         [[Kappa].sub.2] =. 17) and regional
                       disturbances ([[Kappa].sub.1] =-.03,
                       [[Kappa].sub.2]=.26) to be unreliable

Spratt et al(32)     Intraclass correlation
(1990)                 coefficients for judging
                       behavioral responses to
                       NOS ranged from .78
                       to .97

(a) Workers' Compensation Board.

(b) Type of correction used not documented.

(c) LBP=low back pain.

The validity of nonorganic signs for predicting treatment outcomes has been examined by several investigators. Table 4 summarizes the studies that address outcome prediction. Evidence can be found that supports the validity of nonorganic signs for predicting outcomes.[29,31,33,46,47] whereas other researchers question the predictive validity of nonorganic signs.[31,33,46-53] Often, contradictory findings are reported for the same study.[31,33,46,47,50] By definition, some of the signs are scored positively if the patient's pattern of pain is diffuse rather than following a radicular pattern. Patients whose neurological symptoms improve have fewer nonorganic signs.[38,46] These findings suggest that nonorganic signs are more likely to be positive for the patient with mechanical back pain or a diffuse soft tissue pattern of pain. Other studies[29,48,49] suggest that, rather than being a predictor of outcomes, nonorganic signs may change as a result of treatment.

[TABULAR DATA 4 NOT REPRODUCIBLE IN ASCII]

In summary, the nonorganic signs were not developed for the purpose of detecting sincerity of effort, and there is no evidence in the literature to suggest that they can be used for this purpose. Several researchers have raised questions regarding the reliability of the measurements. The evidence regarding the validity of nonorganic signs for predicting outcomes, such as response to treatment and return to work, is inconclusive.[31,33,46-53] In our view, nonorganic signs should be used only for the purpose for which they were originally intended, that is, to identify those persons whose physical recovery may be affected by psychosocial factors and who may need psychosocial assessment or even treatment. Statements regarding sincerity of effort cannot be supported through clinical application of nonorganic signs. Caution should be exercised when using nonorganic signs to predict outcomes such as response to treatment and return to work.

Documentation of Pain Behavior

In reports of musculoskeletal evaluation or FCE of patients with LBP, clinicians frequently document that patients have "exaggerated" or "excessive" pain behavior.[6,9,11,13-15,17,18,24] As discussed in the introduction, these terms add little to our understanding of self-limiting behavior, and they negatively label patients.[26] Research indicates that an individual's pain behavior may be influenced by a variety of factors: environmental,[38] verbal reinforcement,[39] ethnicity,[41] and interaction with one's spouse.[42] Despite these complexities, some clinicians believe that observing and reporting pain behavior has important implications for treatment and prognosis in patients with LBP. The questions clinicians must ask themselves are: (1) Do we have systematic and reliable methods of documenting pain behavior? (2) Should self-reports of pain intensity correlate with pain behavior? and (3) Is the incidence or type of pain behavior related to treatment outcomes or return to work?

An in-depth review of all the methods used to quantify pain behavior is beyond the scope of this article. The reader is referred to a recent review by Solomon[54] for a more thorough treatise on this topic. In Table 5, we briefly summarize the major approaches to quantification of pain behaviors that are pertinent to the patient with LBP. Typically, these methods rely on both verbal cues (eg, vocal complaints, moaning, sighing) and nonverbal cues (eg, bracing, guarding, rubbing, grimacing, posture, use of assistive devices) that are observed while the patient performs a structured set of tasks or during an interview. Some of the methods rely on videotaping,[38,55,56] whereas other methods are designed to be administered "live."[57-59] Evaluation time ranges from 5 minutes of observation[57] to 45 to 60 minutes of interviews.[59] Most of the methods have been studied for reliability,[38,55,57,58,60] with the highest levels of reliability being reported by Keefe and Block[55] and Richards et al.[57] Some of the instruments have been studied for construct validity.[55,56,58,60-64] Some methods have been found to discriminate between patients with pain and control subjects[55,56,60-64] or to be inversely correlated to functional activity level[57] and wellness behaviors.[58] Keefe and Block's method has been validated for both patients with chronic LBP[55,65-67] and patients with "nonchronic" LBP.[68] Several instruments, developed to evaluate pain behavior, have not undergone reliability or validity testing.[38,59,69]

[TABULAR DATA 5 NOT REPRODUCIBLE IN ASCII]

When pain behaviors are inconsistent with self-reports of pain, the patient often is viewed with suspicion.[54] The literature addressing the correlation between observed pain behaviors and self-reports of pain intensity, however, is contradictory. Some investigators[57,59] have found little correlation between these 2 measures, whereas other investigators[55,64] have reported a high correlation between these measures. These seemingly contradictory findings may be explained by the timing and context of the assessments, the type of behavior assessed,[70] the nature of the diagnoses, the method of assessment used, environmental characteristics,[38] and even the physical appearance of the patient.[71] When discrepancies between pain intensity and pain behaviors arise, one measure does not necessarily invalidate the other measure. Instead, we believe these discrepancies should serve as a catalyst for further evaluation that will assist in directing treatment.[70]

The relationship between overt pain behaviors and treatment outcomes also is not well established. Hasenbring et al[72] found that pain behaviors in patients with acute lumbar disk prolapse or protrusion predicted persistent pain 6 months after injury but did not predict application for early retirement. In a sample of 17 patients with chronic LBP, pain behavior during the first epidural block procedure predicted treatment outcomes.[73] The investigators suggested, however, that these results be viewed with caution due to the small sample size. Devulder et al[74] found no correlation between treatment outcomes and pain behavior in patients treated with epidural injections. Hazard et al[75] found no correlation between disability exaggeration and work status. Kleinke and Spangler[76] found no correlation between pain behaviors and other measures of outcome.

In persons with chronic LBP, pain behavior has been noted to improve with treatments that directly address behavioral issues, such as cognitive behavioral therapy (CBT).[77,78] Linton et al[77] combined a cognitive behavioral approach and exercise as a secondary preventive approach for 66 nurses who had sustained a back injury. They found that the intervention group decreased the number of pain behaviors observed during a 10-minute standardized activity session by half, whereas a waiting-list control group increased the number of pain behaviors. The intervention group showed significantly less pain behavior after treatment than did the control group (P=.007). Researchers also have demonstrated successful treatment with CBT in patients with rheumatoid arthritis[62] and in patients with fibromyalgia syndrome.[79] In contrast, pain behavior did not respond to treatment with CBT in patients with osteoarthritis of the hip and knee.[80] With other treatments, such as electrical stimulation implants[81] and drug treatment,[82] pain behaviors do not appear to change, despite improvements in self-reports of pain.

In summary, pain behavior is only one aspect of the complex experience of pain.[26] Failure to include measures that address other cognitive and psychosocial variables that affect the pain experience may result in an incomplete assessment and inappropriate treatment of patients.[26,70] Informal, nonstandardized descriptions of pain during routine clinical practice are subject to considerable error and bias.[83] A variety of pain behavior scales can be administered in a brief period, and some scales have clearly established psychometric properties.[55,57,56,65-68] Clinicians should choose instruments that yield measurements with the highest levels of reliability and validity, consistent with the Standards for Tests and Measurements in Physical Therapy Practice.[22] Still unknown, however, is the correlation of these measures to sincerity of effort or treatment outcomes. Further research is needed before the pain behavior scales can be used for these predictions. Care should be taken, therefore, when interpreting the results of pain behavior scales. Given our current knowledge, pain behavior scales should not be expected to correlate with reports of pain intensity or with outcomes. Their most appropriate use is for documentation of progress toward behavioral goals during rehabilitation. If patients cite pain as a reason for giving a submaximal effort on a majority of tasks of a functional assessment, a comprehensive pain assessment, which is beyond the scope of this publication, is recommended.[26]

The Use of the Coefficient of Variation in Muscle Performance Tests

The coefficient of variation (CV) is a measure of relative standard deviation, in which the standard deviation is divided by the mean and multiplied by 100 to express the standard deviation as a percentage of the mean.[84] In clinical practice, patients sometimes are asked to perform a maximal voluntary contraction repeatedly while using a strength testing device. As reported by Simonsen, "The ratio of the standard deviation divided by the mean is multiplied by 100 to yield a unitless percentage."[85](p516) This measure frequently is used by clinicians to determine whether a patient is giving a consistent effort during testing,[2,7-9,11,16-18] which is then interpreted as a measure of sincerity of effort. The CV is not a statistic that accurately reflects reliability.[23] Instead, the CV describes the variability within a sample, some component of which is measurement error and some component of which is variability among subjects.[23] The assumption is that an intentionally submaximal effort will result in greater variability than a maximal effort.

Several problems, however, exist with the use of the CV in isometric force testing. First, there seems to be a great deal of variability in the CVs reported by different investigators and considerable discrepancy as to exactly how much variability is acceptable.[86-89] Coefficients of variation ranging from 5.1% to 29% have been reported in the peer-reviewed literature.[86-89] Using the CV to determine maximal effort requires that there be a threshold or cutoff point above which the contractions are considered submaximal. The wide variability in reported CVs makes establishing this threshold difficult. Even when submaximal efforts produce higher CVs than those produced by maximal efforts, investigators have failed to establish a definitive threshold or cutoff point, above which the effort could be classified as submaximal.[90,91]

Even if a threshold could be established, it cannot be assumed that patients who give an inconsistent effort are consciously trying to do so. High CVs can occur due to a variety of reasons, such as the type of muscle contraction used, the presence of impairment or pain, the testing protocol or equipment, or the magnitude of the variable in question.[85] Steiner et al[92] reported higher CVs for patients with painful knee syndromes than for subjects without pain when testing eccentric isokinetic knee flexion. Tornvall[93] found that the more joints and muscle groups involved in isometric testing, the higher the CVs. Coefficients of variation for lifting tasks and grip strength, therefore, may be higher than those for elbow or knee flexion. When hand-held dynamometry is used, the examiner's variability may be indistinguishable from the patient's variability.[85] In addition, the magnitude of the variable in question may influence the magnitude of the CV. Thus, individuals who are capable of producing greater force during isometric or isoinertial force testing will have a lower CV than an individual who is capable of exerting less force, merely due to the nature of the statistic itself.[23]

Based on a review of 88 English-language, peer-reviewed studies published before March 1992, Newton and Waddell concluded that "there is no evidence that [using] iso-machines [to determine CVs] provides a reliable or valid method to assess effort or to detect if the person is faking."[4](p808) Studies cited in this review demonstrated no differences in CVs between maximal and submaximal efforts[94] and showed that subjects could reliably repeat a submaximal effort when performing isometric contractions for plantar flexion, hip flexion, and a lilting task.[95] Subsequent to this review, Newton et al[96] found measures of consistency of effort in isokinetic isoinertial testing to be unreliable and incapable of distinguishing between maximal and submaximal efforts. Simonsen[85] examined the correlation between CVs for static tasks of the ERGOS Work Simulator.(*) He found differences in the mean CVs between tasks and concluded that CVs will vary depending on the tasks being tested. These studies[85,94-96] and the cited review[4] provide convincing evidence that the CV cannot be used to determine sincerity of effort.

In summary, use of the CV to measure sincerity of effort is unsubstantiated in the literature. Although there is evidence that submaximal efforts can be reliably reproduced, measures of CV vary greatly depending on the instrumentation used, the task performed, the muscle groups tested, and the presence of pain. Additionally, there is little agreement between investigators and clinicians regarding the appropriate threshold CV for determining submaximal effort. Finally, there is little, if any, theoretical basis for using the CV as an index of reliability.[23] For these reasons, the CV should not be used to determine sincerity of effort.

Correlation Between Musculoskeletal Evaluation and FCE

A lack of correlation between measures of impairment determined during musculoskeletal evaluations and more functional measures is sometimes used as evidence that the patient is giving less than full effort during testing.[17] Research, however, suggests that impairment is not directly (ie, linearly) correlated with function.[97-99] Range-of-motion (ROM) measures have been found to correlate only moderately with measures of function.[100,101] Lankhorst et al[102] found that muscle force production is not correlated with function. Newton et al[96] found that isokinetic trunk muscle performance was poorly correlated with measures of disability. Roberson et al[103] found that measures of isometric and isoinertial trunk extensor performance cannot be used to predict lilting ability. MacKenzie et al[104] found low correlations between both lower-extremity ROM and muscle force and function at work and at home in patients with lower-extremity fractures. Waddell et al[105] studied patients with chronic LBP and found that a combination of measures of spinal ROM, straight leg raising, spinal tenderness, and sit-ups explained only 25% of the variance of disability.

In summary, there are no studies confirming a direct (ie, linear) relationship between musculoskeletal impairments and function. As with CVs, further research would need to establish the relationship between impairment and function as well as the thresholds or cutoff points that can be used for determining sincerity of effort.

Grip Measures

Three approaches to documenting sincerity of effort using measures of isometric grip force have evolved and are widely used: (1) calculation of the CV of repeated measures,[2,8,9,11,12,16,17] (2) analysis of force-handle position curves (bell-shaped curve),[7,11,13,17,20,21] and (3) comparison between slow sustained measures and rapid assessment measures, known as the rapid exchange grip (REG).[106] Other, less widely used methods include (1) force-time curves,[107,108] (2) rapid simultaneous grip (RSG),[109] and (3) ratio of peak and average force and ratio of slope and peak force.[108,110] Table 6 summarizes the studies that address the methods for detecting sincerity of effort with grip force.

[TABULAR DATA 6 NOT REPRODUCIBLE IN ASCII]

The idea that measures of isometric grip force could be used to determine sincerity of effort was first suggested by Bechtol[111] in his initial report describing a grip dynamometer with adjustable handle spacing to accommodate a variety of hand sizes. An important aspect of this study was the establishment of the relationship of force generation and handle position and the variables that affected consistency in testing grip force. Mathiowetz et al[112] later standardized the test position (subjects seated with shoulder adducted and neutrally rotated, elbow flexed to 90 [degrees], forearm in neutral, and wrist in 0 [degrees]-30 [degrees] of extension and 0 [degrees]-15 [degrees] of ulnar deviation) and procedure (3 successive trials recorded with dynamometer set to the second handle position). They also discussed reliability and validity but did not address the issue of detecting sincerity of effort.

The CV has been discussed as it relates to isometric force testing of the trunk and extremities. Further discussion of the validity of the CV in grip tests, however, seems warranted due to its wide clinical use.[2,8,9,11,12,16,17] Despite the widespread use of this approach, only Robinson et al[2] have addressed the scientific value of this protocol. In this study, both submaximal and maximal efforts demonstrated low intraclass correlation coefficients, indicating poor stability of these measures. Robinson et al also found that a high percentage of the submaximal efforts were incorrectly classified as maximal, indicating that this approach has a high percentage of false negatives. These investigators concluded that "using individual CV measures of effort consistency is not to be recommended."[2](p49)

The bell-shaped curve approach was developed by Stokes.[113] He found that patients he considered to have a "true" hand injury showed a pattern over the dynamometer's 5 handle positions, with the greatest force exerted at the middle handle position and less force exerted at the wider and narrower positions. In contrast, he found that patients he suspected of giving an insincere effort demonstrated equal force at each of the 5 handle positions, yielding a fiat, horizontal linear relationship. One problem with his study, however, was that Stokes provided no evidence of absence of disease in patients he judged as giving an insincere effort. Factors such as pain could have influenced the ability of the patients to produce more force at the middle handle position.

In a later study, Stokes at al,[114] studied subjects without pain who were asked to exert sincere and insincere efforts and patients with hand pain who were considered sincere and insincere. Both insincere groups (subjects without pain and patients) were found to have lower variability in force production at the 5 handle positions (flatter bell-shaped curve). One problem with both studies conducted by Stokes and colleagues[113,114] is that no objective process for identifying "insincere" patients was reported. Niebuhr and Marion[115,116] subsequently studied the bell-shaped curve in subjects without pain who were instructed to exert true and feigned maximal efforts, but they were unable to replicate the findings of Stokes and colleagues.[113,114]

In current clinical application, the judgment of whether the curve has an adequate bell shape and, therefore, whether a person is exerting a sincere effort or an insincere effort, typically is made by visual observation of the curve. This judgment, therefore, is nothing more than an individual clinician's opinion. The research addressing the value of using the bell-shaped curve for detecting sincerity of effort was conducted using complex statistical trend analysis techniques that few clinicians have at their disposal.

The force-time curve is a less well known approach to the detection of sincerity of effort and was first described by Smith et al.[107] Subjects are asked to sustain a maximal grip for 5 seconds. In a study of subjects without pain using the Smith protocol,[108] force-time curves for sincere-effort trials demonstrated an initial rapid rise in force that was sustained over 5 seconds. In insincere efforts, the initial rapid rise in force was followed by a "relatively gradual decline over the last few seconds."[108(p150)] Subjects who gave sincere efforts were found to have less deviation between peak force and average force compared with subjects who gave insincere efforts. A 2-factor analysis of variance revealed differences in ratios of peak and average force production between sincere and insincere efforts. The sensitivity tot discriminating between sincere and insincere efforts was 90%, and the specificity for discriminating between sincere and insincere efforts was 85%. As with the bell-shaped curve, the discriminative ability of the force-time curve was not achieved using visual analysis of the data. Instead, statistical applications were used. In order for this approach to be useful and efficient for the clinicians, access to trend-analysis software would be necessary.

The REG and RSG tests are based on the premise that individuals have more difficulty maintaining a submaximal effort when the speed of grip force testing is increased from 1 sustained squeeze to 80 to 90 squeezes per minute. Those individuals who are not giving a sincere effort are supposed to show higher forces with REG testing than with slow sustained grip force testing. Consensus on the threshold for insincere effort, however, has not been achieved. Early test developers[106,117] had very vague criteria for determining a positive test versus a negative test. Czitrom and Lister,[118] for example, defined a positive REG as one in which the REG is "dramatically higher" than static grip measures. No operational definition or data were provided to quantify a specific cutoff point. By comparison, Hildreth et al[117] found a 79% greater REG in subjects who were instructed to feign injury, whereas Joughin et al[109] considered a 25% increase in REG over sustained grip force to be "positive." In addition, extremely high standard deviations for REG ([+ or -] 60%) were found by Hildreth et al.[117]

Joughin et al[109] modified the REG protocol such that rapid gripping of the dynamometer was performed simultaneously with both hands. The premise of this modification, termed the "rapid simultaneous grip," was that bilateral gripping made it even more difficult to differentiate the performance of each hand. These investigators determined cutoff points based on the calculated sensitivities and specificities of each test. They recommended different thresholds for patients with hand impairments and for subjects without hand impairments (Tab. 6).

The ratios of average and peak force, peak force and body weight, and slope and peak force were studied by Chengalur et al[108] and Gilbert and Knowlton.[110] They investigated the effectiveness of the ratios of slope and peak force, average and peak force, and peak force and body weight for determining sincere versus insincere efforts. Chengalur et al[108] found differences in ratios of average and peak force between subjects who gave sincere efforts and subjects who gave insincere efforts. The magnitude of the differences between sincere and "faking" trials for these ratios ranged from 0.82 to 66.86. The accuracy for predicting true sincere trials ranged from 90.0% to 96.7%, whereas the accuracy for predicting true "faking" trials ranged from 22.5% to 76.7%. Gilbert and Knowlton[110] found that for female subjects, the ratio of average and peak force was the only discriminator between subjects who gave sincere efforts and subjects who gave insincere efforts, correctly classifying 94% of the subjects. The ratio of average and peak torque and the ratio of slope and peak torque were found to correctly classify 85% of the male subjects. Their calculations of test-retest reliability yielded correlation coefficients ranging from .80 to .95 for subjects who gave insincere efforts, which suggests that these subjects can consistently reproduce submaximal efforts.

None of the studies addressing the use of grip force measures in detecting sincerity of effort were conducted with patients with LBP. The validity of measurements obtained with these methods for these patients, therefore, is unknown. Studies addressing the validity of measurements obtained with a variety of grip force measures for detecting sincerity of effort in persons with and without hand dysfunction further suggest that none of these measures have been validated adequately for this purpose.[1] Coefficients of variation calculated over 3 grip force trials are unstable and produce a high percentage of false negatives. Studies addressing the validity of the bell-shape curve approach are contradictory. The research surrounding REG is controversial regarding the cutoff scores for identifying sincere versus insincere efforts. Force-time curves, ratios of average and peak torque, and ratios of slope and peak torque are valid only when they are analyzed statistically, using techniques not currently used by most clinicians. Assessment of sincerity of effort using force-time or bell-shape curves is not valid when based on visual assessment of the data. Clinicians are advised to avoid using the CV, REG, and bell-shaped curve approaches for detecting sincerity of effort, as the literature does not support the reliability and validity of their measurements for this purpose. If force-time curves or ratios between peak force, average force, and slope are used, the results should be interpreted with caution and analyzed statistically as described by the test developers.

Relationship of Heart Rate to Pain Intensity

Another method used to determine the sincerity of effort during FCEs is to compare ratings of pain intensity on a visual analog scale (VAS) or verbal ratings with the heart rate. The premise is that as pain increases, the heart rate also increases, and when patients report high pain scores without concomitantly high heart rates, they are consciously trying to exaggerate their pain. Research conducted by Borg et al[119] is sometimes cited to justify this approach. Careful examination of Borg and colleagues' work, however, leads to questions about this interpretation. The relationship between heart rate and pain in this study was established for patients with angina pectoris exercising with a very specific bicycle ergometry protocol. The ischemic pain of angina pectoris and its associated cardiovascular consequences are somatically and physiologically different from musculoskeletal pain. The physiologic relationship among chest pain, heart rate, and workload are not the same as the relationship among LBP, heart rate, and workload. The results of Borg's studies, therefore, cannot be extrapolated to patients with musculoskeletal dysfunction.

Other studies[120-122] have addressed the heart rate response to acutely painful stimuli. The researchers found mild transient increases in heart rate in response to painful heat or cold stimuli. Coghill et al[122] found that when a painful stimulus is merely anticipated, the anxiety alone may increase heart rate. This study has important implications for correlations between pain ratings and heart rate because, in these authors' clinical experience, most of the individuals with LBP are anxious in anticipation of pain during examination.

When interpreting the results of these studies, clinicians should be aware that (1) they involved individuals without pain, (2) the painful stimulus, either heat or cold, was applied with rapid onset and rapid cessation, (3) the heart rate increases, although statistically significant, were small (5-10 beats per minute) and transient (lasting only seconds), and (4) the heart rate monitors used were more accurate and sophisticated devices than the typical chest strap models that are commonly used in physical therapy clinics. For these reasons, generalization of the results of these studies to persons with LBP is limited.

Only one group of researchers addressed the physiological response to pain in patients with chronic pain. Peters and Schmidt[123] studied physiological responses to repeated acutely painful mechanical stimuli in persons with and without chronic LBP. In response to the painful stimuli, both groups had increased skin conductance fluctuations and respiratory rates, but they did not demonstrate increases in heart rate. No correlation was found between the responses to painful stimuli and the heart and respiratory rates in either the subjects with chronic LBP or the subjects without chronic LBP. These findings suggest that heart rate cannot be used to validate self-report measures of pain.

In summary, using autonomically mediated physiological measures to validate self-report measures of pain is attractive to therapists who seek objective measures of pain. Studies supporting this relationship in individuals without LBP receiving an acutely painful thermal stimulus cannot be generalized to individuals with LBP. A relationship between a mechanical painful stimulus and heart rate has not been found in patients with chronic LBP. Preliminary research suggests a complex and poorly understood relationship among pain intensity, physiologic responses, and pain perception.[120-123] Based on these studies, accusing patients of exaggerating their pain response due to a lack of concomitant rise in heart rate is not appropriate.

Clinical Implications

The very concept or construct of sincerity of effort is illusive and difficult to measure. The definition explicitly implies a measurement of motivation. To date, none of the previously discussed methods for detecting sincerity of effort have been adequately studied for its use with patients with LBP. The medicolegal and ethical implications of this lack of validation are tremendous. Clinical reports that imply the patient has intentionally given less than a full effort are in clear violation of American Physical Therapy Association (APTA) measurement standards[22] and often have extremely negative consequences for patients with LBP. Patients may have Workers' Compensation payments withheld, may lose their jobs, may receive a diminished medicolegal settlement, or may undergo inappropriate clinical treatment as a result of the negative labeling associated with the approaches discussed. Any methods used to make determinations regarding sincerity of effort, therefore, should have research published in the peer-reviewed literature to support the reliability and validity of their measurements for this purpose.

The 1993 US Supreme Court ruling of Daubert v Merill-Dow Pharmaceuticals addressed the legal implications of judgments of sincerity of effort. This ruling established the criteria for admissibility of evidence in courts of law. According to this ruling, expert witness testimony must be based on measurements that have established reliability and validity published in peer-reviewed journals.[124] Given the available literature, the reviewed approaches for documenting sincerity of effort do not meet the criteria established by this ruling,[124] nor do they comply with APTA measurement standards.[22]

No approach for detecting sincerity of effort has been directly correlated to outcome. If patients demonstrate a "high" CV, if they display a large number of pain behaviors, if they score positively on the nonorganic signs test, if their impairment measures do not correlate with function, or if their heart rate does not increase with pain reports, these findings do not necessarily predict poor adherence or a failure to return to work. Unfortunately, any of these statements can negatively bias health care professionals against the patient. Those patients who are negatively labeled may receive inadequate or inappropriate treatment.

The question remains: Do we need to measure sincerity of effort? If we ignore this issue, then some people may receive treatment they do not need, may fail to return to appropriate work, or may receive undeserved disability payments, all of which are costly to industry and society. If we make judgments with the currently available methods, we are taking a great risk of incorrectly classifying some patients as insincere. This misclassification may cost them their job or their medical or Workers' Compensation insurance and may negatively affect their self-esteem. As a profession, we are under pressure from referral sources to assess sincerity of effort. We seriously question whether detecting sincerity of effort is an appropriate role for clinicians.

The APTA standards for measurement and practice[22] require that clinical measurements used to detect sincerity of effort have established validity. Currently, clinicians do not have legitimate tools or methods with which to make these assessments. Any statements regarding sincerity of effort, therefore, are strictly clinical opinion. Therapists are advised to avoid the use of the discussed methods for the purpose of supporting claims of detecting sincerity or level of cooperation with evaluation. Therapists who draw unwarranted conclusions from test results are violating the rights of the person being tested.[22] Therapists also are advised to avoid reporting test results as "valid" or "invalid" based on perceived levels of cooperation and to avoid using the terms "symptom magnification" and "exaggerated pain behavior" to describe patient behavior.

We suggest, instead, that clinicians and referral sources seek alternative methods to address delayed recovery and to understand the biobehavioral factors affecting pain and disability (ie, disease conviction, perceptual-cognitive bias, perceived control, perceived disability, fear of pain, perceptions of work and family, and self-efficacy).[26] Feuerstein and Beattie[26] provided an excellent theoretical framework that can enhance our understanding of these factors. In addition, these authors reviewed some commonly used approaches for assessment of biobehavioral factors and provided examples of ways in which physical therapists can use this information in clinical practice. Further research is needed to identify the most appropriate measures and treatments and to understand their implications for clinical practice. Identifying and addressing the biobehavioral factors that affect delayed recovery, however, may provide a more proactive approach to achieving functional restoration.[26]

(*) Work Recovery Inc, 2341 S Friebus, Suite 14, Tucson, AZ 85713.

References

[1] King PM. Analysis of approaches to detection of sincerity of effort through grip strength measurement. Work. 1998;10:9-13.

[2] Robinson ME, Geisser ME, Hanson CS, O'Conner PD. Detecting submaximal efforts in grip strength testing with the coefficient of variation. Journal of Occupational Rehabilitation. 1993;3:45-50.

[3] Baker JC. Burden of proof in detection of submaximal effort. Work. 1998;10:63-70.

[4] Newton M, Waddell G. Trunk testing with isomachines, part 1: review of a decade of scientific evidence. Spine. 1993;18:801-811.

[5] Alpert J. State of California Functional Capacity Evaluation Guidelines for Physical Therapists. Sacramento, Calif: California Chapter, American Physical Therapy Association; 1996:5, 6, 8.

[6] Bottomly S. Interpretive Manual to Accompany the Residual Functional Evaluation for Unscheduled Disability. Salem, Ore: Work Hardening Association of Oregon in conjunction with the Workers' Compensation Division, State of Oregon Department of Insurance and Finance; 1989:12-16.

[7] Oregon Administrative Rules for Department of Consumer and Business Services, Workers' Compensation Division. Salem, Ore: 1997. Chapter 436, Div 010-32.

[8] Matheson LN. Functional capacity evaluation. In: Demeter SL, Andersson GBJ, Smith GM, eds. Disability Evaluation. St Louis, Mo: Mosby; 1996:180.

[9] Matheson LN. Use of the BTE Work Simulator to screen for symptom magnification syndrome. Industrial Rehabilitation Quarterly. 1989; 2(2):5-31.

[10] Key GL. Work conditioning and work hardening. In: Key GL, ed. Industrial Therapy. St Louis, Mo: Mosby; 1995:254-294.

[11] Blankenship K. The Blankenship System Functional Capacity Evaluation: The Procedure Manual. Macon, Ga: The Blankenship Corp; 1994: 4.01-4.83, 5.01-5.17, 9.01-9.37.

[12] Jacobs J. BTE Clinical Applications Manual: A Clinically Oriented Reference Manual for Users of the BTE Work Simulator. Baltimore, Md: BTE Co; 1992:3-1-3-12.

[13] Saunders RL. Industrial Rehabilitation: Techniques for Success. Chaska, Minn: The Saunders Group; 1995:34-38.

[14] Saunders HD, Saunders RL. Evaluation, Treatment, and Prevention of Musculoskeletal Disorders, Volume 1: Spine. Chaska, Minn: The Saunders Group; 1993:44, 96, 358.

[15] Saunders RL, Beissner KL, McManis BG. Estimates of weight that subjects can lift frequently in functional capacity evaluations. Phys Ther. 1997;77:1717-1728.

[16] ERGOS Work Simulator Administration Manual. Tucson, Ariz: Work Recovery Inc; 1987: Interims 8.

[17] Isernhagen SJ. Isernhagen Work Systems Functional Capacity Evaluation. Duluth, Minn: Isernhagen Work Systems; 1996:52a-52b, 83.

[18] Tatom AJ. Dynametric testing of the trunk. In: Basmajian JV, Nyberg R, eds. Rational Manual Therapies. Baltimore, Md: Williams & Wilkins; 1993:173.

[19] Ogden-Niemeyer L. The issue of abnormal illness behavior in work hardening. In: Ogden-Niemeyer L, Jacobs K, eds. Work Hardening: State of the Art. Thorofare, NJ: Slack Inc; 1989:83-109.

[20] Magee DJ. Orthopedic Physical Assessment. Philadelphia, Pa: WB Saunders Co; 1997:297.

[21] Bear-Lehman J. Functional evaluation of the wrist and hand. In: Nordin M, Andersson GBJ, Pope MH, eds. Musculoskeletal Disorders in the Workplace. St Louis, Mo: Mosby; 1997:444, 452.

[22] Task Force on Standards for Measurement in Physical Therapy. Standards for tests and measurements in physical therapy practice. Phys Ther. 1991;71:589-622.

[23] Rothstein JM, Echternach JL. Primer on Measurement: An Introductory Guide to Measurement Issues. Alexandria, Va: American Physical Therapy Association; 1993.

[24] Monsein M, Clift RB. Pain and return to work: turning the corner. In: Isernhagen SJ, ed. The Comprehensive Guide to Work Injury Management. Gaithersburg, Md: Aspen Publishers Inc; 1995:546.

[25] Dirckx JH, ed. Stedman's Concise Medical Dictionary for the Health Professions. 3rd ed. Baltimore, Md: Williams & Wilkins; 1997:845.

[26] Feuerstein M, Beattie PF. Biobehavioral factors affecting pain and disability in low back pain: mechanisms and assessment. Phys Ther. 1995;75:267-280.

[27] Lechner DE, Jackson JR, Roth DL, Straaton KV. Reliability and validity of a newly developed test of physical work performance. J Occup Med. 1994;36:997-1004.

[28] Landis JR, Koch GC. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159-174.

[29] Waddell G, McCulloch JA, Kummel E, Venner RM. Nonorganic physical signs in low-back pain. Spine. 1980;5:117-125.

[30] Scalzitti DA. Screening for psychological factors in patients with low back problems: Waddell's nonorganic signs. Phys Ther. 1997;77: 306-312.

[31] Lehmann TR, Russell DW, Spratt KF. The impact of patients with nonorganic physical findings on a controlled trial of transcutaneous electrical nerve stimulation and electroacupuncture. Spine. 1983;8: 625-634.

[32] Spratt KF, Lehmann TR, Weinstein JN, Sayre HA. A new approach to the low-back physical examination: behavioral assessment of mechanical signs. Spine. 1990;15:96-102.

[33] Doxey NC, Mitson GL, Dzioba RB, Lacroix JM. Predictors of outcome in back surgery candidates. J Clin Psychol. 1988;44:611-621.

[34] Mooney V, Robertson J. The facet syndrome. Clin Orthop. 1976;115: 149-156.

[35] Badgley CE. The articular facets in relation to low-back pain and sciatic radiation. J Bone Joint stag. 1941;23:481-496.

[36] Hirsch C, Ingelmark BE, Miller M. The anatomical basis for low back pain: studies on the presence of sensory nerve endings in ligamentous, capsular, and intervertebral disc structures in the human lumbar spine. Acta Orthop Scand. 1963;33:1-17.

[37] Inman VT, Saunders JB. Referred pain from skeletal structures. J Nerv Ment Dis. 1944;99:660-667.

[38] Cinciripini PM, Floreen A. An assessment of chronic pain behaviour in a structured interview. J Psychosom Res. 1983;27:117-123.

[39] Cairnes D, Pasina J. Comparison of verbal reinforcement and feedback in the operant treatment of disability due to chronic low back pain. Behavioral Therapy. 1977;8:621-630.

[40] Waddell G, Birchner M, Finlayson D, Main CJ. Symptoms and signs: Physical illness or illness behavior? BMJ. 1984;289:739-741.

[41] Tursky B, Sternbach RA. Further physiological correlates of ethnic differences in responses to shock. Psychophysiology. 1967:4:67-74.

[42] Block AR, Kremer EF, Gaylor M. Behavioral treatment of chronic pain: the spouse as a discriminative cue for pain behavior. Pain. 1980;9:243-252.

[43] Korbon GA, DeGood DE, Schroeder ME, et al. The development of a somatic amplification rating scale for low-back pain. Spine. 1987;12: 787-795.

[44] Waddell G, Main CJ, Morris EW, et al. Normality and reliability in the clinical assessment of backache. Br Med J Clin Res. 1982;284: 1519-1523.

[45] McCombe PF, Fairbank JCT, Cockersole BC, Pynsent PB. Reproducibility of physical signs in low-back pain. Spine. 1989; 14:908-918.

[46] Dzioba RB, Doxey NC. A prospective investigation into the orthopaedic and psychologic predictors of outcome of first lumbar surgery following industrial injury. Spine. 1984;9:614-623.

[47] Karas R, McIntosh G, Hall H, et al. The relationship between nonorganic signs and centralization of symptoms in the prediction of return to work for patients with low back pain. Phys Ther. 1997;77: 354-360.

[48] Waddell G, Main CJ, Morris EW, et al. Chronic low-back pain, psychologic distress, and illness behavior. Spine. 1984;9:209-213.

[49] Waddell G, Morris EW, Di Paola MP, et al. A concept of illness tested as an improved basis for surgical decisions in low-back disorders. Spine. 1986; 11:712-719.

[50] Lancort J, Kettelhut M. Predicting return to work for lower back pain patients receiving workers' compensation. Spine. 1992;17: 629-640.

[51] Bradish CF, Lloyd GJ, Aldam CH, et al. Do nonorganic signs help to predict the return to activity of patients with low-back pain? Spine. 1988;13:557-560.

[52] Chan CW, Goldman S, Ilstrup DM, et al. The pain drawing and Waddell's nonorganic physical signs in chronic low-back pain. Spine. 1993;18:1717-1722.

[53] Lacroix JM, Powell J, Lloyd GJ, et al. Low-back pain: factors of value in predicting outcome. Spine. 1990;15:495-499.

[54] Solomon PE. Measurement of pain behaviour. Physiotherapy Canada. 1996;48:52-58.

[55] Keefe FJ, Block AR. Development of an observation method for assessing pain behavior in chronic low back pain patients. Behavioral Therapy. 1982;13:363-375.

[56] Follick M, Ahern D, Aberger E. Development of an audiovisual taxonomy of pain behaviour: reliability and discriminant validity. Health Psychol. 1985;4:555-568.

[57] Richards JS, Nepomuceno C, Riles M, Suer Z. Assessing pain behavior: the UAB Pain Behavior Scale. Pain. 1982;14:393-398.

[58] Vlaeyen JWS, Pernot DFM, Kole-Snijders AMJ, et al. Assessment of the components of observed chronic pain behavior: the Checklist for Interpersonal Pain Behavior (CHIP). Pain. 1990;43:337-347.

[59] Dirks JF, Wunder J, Kinsman R, et al. A pain rating scale and a pain behavior checklist for clinical use: development, norms, and the consistency score. Psychother Psychosom. 1993;59:41-49.

[60] Anderson KO, Bradley LA, McDaniel LK, et al. The assessment of pain in rheumatoid arthritis: disease differentiation and temporal stability of a behavioral observation method. J Rheumatol. 1987;14: 700-704.

[61] Anderson KO, Bradley LA, McDaniel LK, et al. The assessment of pain in rheumatoid arthritis: validity of a behavioral observation method. Arthritis Rheum. 1987;30:36-43.

[62] Bradley LA, Young LD, Anderson KO, et al. Effects of psychological therapy on pain behavior of rheumatoid arthritis patients: treatment outcome and six-month follow-up. Arthritis Rheum. 1987;30:1105-1114.

[63] Jaworski TM, Bradley LA, Heck LW, et al. Development of an observation method for assessing pain behaviors in children with juvenile rheumatoid arthritis. Arthritis Rheum. 1995;38:1142-1151.

[64] McDaniel LK, Anderson KO, Bradley LA, et al. Development of an observation method for assessing pain behavior in rheumatoid arthritis patients. Pain. 1986;24:165-184.

[65] Keefe FJ, Wilkins, RH, Cook WA. Direct observation of pain behavior in low back pain patients during physical examination. Pain. 1984;20:59-68.

[66] Keefe FJ, Hill RW. An objective approach to quantifying pain behavior and gait patterns in low back pain patients. Pain. 1985;21: 153-161.

[67] Keefe FJ, Crisson JE, Maltbie A, et al. Illness behavior as a predictor of pain and overt behavior patterns in chronic low back pain patients. J Psychosom Res. 1986;30:543-551.

[68] Jensen IB, Bradley LA, Linton SJ. Validation of an observation method of pain assessment in non-chronic back pain. Pain. 1989;39: 267-274.

[69] Turk DC, Wack JT, Kerns RD. An empirical examination of the "pain-behavior" construct. J Behav Med. 1985;8:119-130.

[70] Turk DC, Flor H. Pain greater than pain behaviors: the utility and limitations of the pain behavior construct. Pain. 1987;31:277-295.

[71] Hadjistavropoulos HD, Ross M, von Baeyer CL. Are physicians' ratings of pain affected by patients' physical attractiveness? Soc Sci Med. 1990;31:69-72.

[72] Hasenbring M, Marienfeld G, Kuhlendahl D, Soyka D. Risk factors of chronicity in lumbar disc patients: a prospective investigation of biologic, psychologic, and social predictors of therapy outcome. Spine. 1994; 19:2759 -2765.

[73] Connolly GH, Sanders SH. Predicting low back pain patients' response to lumbar sympathetic nerve blocks and interdisciplinary rehabilitation: the role of pretreatment overt pain behavior arid cognitive coping strategies. Pain. 1991;44:139-146.

[74] Devulder J, Bogaert L, Castille F, et al. Relevance of epidurography and epidural adhesiolysis in chronic failed back surgery patients. Clin J Pain. 1995;11:147-150.

[75] Hazard RG, Bendix A, Fenwick JW. Disability exaggeration as a predictor of functional restoration outcomes for patients with chronic low-back pain. Spine. 1991;16:1062-1067.

[76] Kleinke C, Spangler A. Psychometric analysis of the audiovisual taxonomy for assessing pain behaviour in chronic low back pain patients. J Behav Med. 1988;11:83-95.

[77] Linton SJ, Bradley LA, Jensen I, et al. The secondary prevention of low back pain: a controlled study with follow-up. Pain. 1989;36: 197-207.

[78] Turner JA. Comparison of group progressive-relaxation training and cognitive-behavioral group therapy for chronic low back pain. J Consult Clin Psychol. 1982;50:757-765.

[79] White KP, Nielson WR. Cognitive behavioral treatment of fibromyalgia syndrome: a follow-up assessment. J Rheumatol. 1995;22:717-721.

[80] Keefe FJ, Caldwell DS, Williams DA, et al. Pain coping skills training in the management of osteoarthritic knee pain: a comparative study. Behavioral Therapy. 1990;21:49-62.

[81] Strege DW, Cooney WP, Wood MB, et al. Chronic peripheral nerve pain treated with direct electrical nerve stimulation. J Hand Surg [Am]. 1994; 19:931-939.

[82] Anderson KO, Bradley LA, Turner RA, et al. Pain behavior of rheumatoid arthritis patients enrolled in experimental drug trials. Arthritis Care Research. 1994;7(2):64-68.

[83] Waddell G, Richardson J. Observation of overt pain behaviour by physicians during routine clinical examination of patients with low back pain. J Psychosom Res. 1992;36:77-87.

[84] Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice. East Norwalk, Conn: Appleton & Lange; 1993:680.

[85] Simonsen JC. Coefficient of variation as a measure of subject effort. Arch Phys Med Rehabil. 1995;76:516-520.

[86] Agre JC, Magness JL, Hull SZ, et al. Strength testing with a portable dynamometer: reliability for upper and lower extremities. Arch Phys Med Rehabil. 1987;68:454-458.

[87] Chaffin DB, Herrin GD, Keyserling WM. Preemployment strength testing: an updated position. J Occup Med. 1978;20:403-408.

[88] Montgomery LC, Douglass LW, Deuster PA. Reliability of an isokinetic test of muscle strength and endurance. J Orthop Sports Phys Ther. 1989;10:315-322.

[89] Kannus P. Normality, variability, and predictability of work, power, and torque acceleration energy with respect to peak torque in isokinetic muscle testing. Int J Sports Med. 1992;13:249-256.

[90] Bohannon RW. Differentiation of maximal from submaximal static elbow flexor efforts by measurement variability. Am J Phys Med. 1987;66:213-218.

[91] Saunders N, Bohannon R. Can feigned maximal efforts be distinguished from maximal efforts? Journal of Human Muscle Performance. 1991;1:16-24.

[92] Steiner LA, Harris BA, Krebs DE. Reliability of eccentric isokinetic knee flexion and extension measurements. Arch Phys Med Rehabil. 1993;74:1327-1335.

[93] Tornvall G. Assessment of physical capabilities. Acta Physiol Scand Suppl. 1963;201:1-102.

[94] Robinson M, Macmillan M, O'Connor P, et al. Reproducibility of maximal versus submaximal efforts in an isometric lumbar extension task. J Spinal Disord. 1991;4:444-448.

[95] Carlsoo S. With what degree of precision can voluntary static muscle force be repeated? Scand J Rehabil Med. 1986;18:1-3.

[96] Newton M, Thow M, Somerville D, et al. Trunk strength testing with iso-machines, part 2: experimental evaluation of the Cybex II back testing system in normal subjects and patients with chronic low back pain. Spine. 1993;18:812-824.

[97] Yelin EH, Henke CJ, Epstein WV. Work disability among persons with musculoskeletal conditions. Arthritis Rheum. 1986;29:1322-1333.

[98] Althouse HL. Preplacement screening: revealing a true profile of musculoskeletal abilities. Occup Health Saf. 1980;49(5):25-30.

[99] Jette AM. Concepts of health and methodological issues in functional status assessment. In: Granger CV, Gresham GE, eds. Functional Assessment in Rehabilitation. Baltimore, Md: Williams & Wilkins; 1984:46-64.

[100] Bergstrom G, Aniansson A, Bejelle A, et al. Functional consequences of joint impairment at age 79. Scand J Rehabil Med. 1985;17: 183-190.

[101] Badley EM, Wagstaff S, Wood PH. Measures of functional ability (disability) in arthritis in relation to impairment of range of joint movement. Ann Rheum Dis. 1984;43:563-569.

[102] Lankhorst GJ, Van de Stadt RJ, Van der Korst JK. The relationships of functional capacity, pain, and isometric and isokinetic torque in osteoarthritis of the knee. Scand J Rehabil Med. 1985; 17:167-172.

[103] Roberson JM, Witt P, Gross MT. A comparison of trunk extensor strength and squat lifting ability. J Orthop Sports Phys Ther. 1997;25:137-144.

[104] MacKenzie EJ, Cushing BM, Jurkovich GJ, et al. Physical impairment and functional outcomes six months after severe lower extremity fractures. J Trauma. 1993;34:528-538.

[105] Waddell G, Somerville D, Henderson I, Newton M. Objective clinical evaluation of physical impairment in chronic low back pain. Spine. 1992;17:617-628.

[106] Lister G. The Hand: Diagnosis and Indications. 3rd ed. New York, NY: Churchill Livingstone Inc; 1993:162-163.

[107] Smith GA, Nelson RC, Sadoff SJ, Sadoff AM. Assessing sincerity of effort in maximal grip strength tests. Am J Phys Med Rehabil. 1989;68:73-80.

[108] Chengalur SN, Smith GA, Nelson RC, Sadoff AM. Assessing sincerity of effort in maximal grip strength tests. Am J Phys Med Rehabil. 1990;69:148-153.

[109] Joughin K, Gulati P, Mackinnon SE, et al. An evaluation of rapid exchange and simultaneous grip tests. J Hand Surg[Am]. 1993;18:24.5-252.

[110] Gilbert JC, Knowlton RG. Simple method to determine sincerity of effort during a maximal isometric test of grip strength. Am J Phys Med. 1983;62:135-144.

[111] Bechtol C. Grip test the use of a dynamometer with adjustable handle spacings. J Bone Joint Surg Am. 1954;36:820-832.

[112] Mathiowetz V, Weber K, Volland G, Kashman N. Reliability and validity of grip and pinch strength evaluations. J Hand Surg [Am]. 1984;9:222-226.

[113] Stokes HM. The seriously uninjured hand: weakness of grip. J Occup Med. 1983;25:683-684.

[114] Stokes HM, Landrieu KW, Domangue B, Kunen S. Identification of low-effort patients through dynamometry. J Hand Surg [Am]. 1995; 20:1047-1056.

[115] Niebuhr BR, Marion R. Detecting sincerity of effort when measuring grip strength. Am J Phys Med Rehabil. 1987;66:16-24.

[116] Niebuhr BR, Marion R. Voluntary control of submaximal grip strength. Am J Phys Med Rehabil. 1990;69:96-101.

[117] Hildreth DH, Breidenbach WC, Lister GD, Hodges AD. Detection of submaximal effort by use of the rapid exchange grip. J Hand Surg [Am]. 1989;14:742-745.

[118] Czitrom AA, Lister GD. Measurement of grip strength in the diagnosis of wrist pain. J Hand Surg [Am]. 1988;13:16-19.

[119] Borg G, Holmgren A, Lindblad I. Quantitative evaluation of chest pain. Acta Med Scand Suppl. 1981;644:43-45.

[120] Moltner A, Holzo R, Strian F. Heart rate changes as an autonomic component of the pain response. Pain. 1990;43:81-89.

[121] Kregel KC, Seals DR, Callister R. Sympathetic nervous system activity during skin cooling in humans: relationship to stimulus intensity and pain sensation. J Physiol (Lond). 1992;454:359-371.

[122] Coghill RC, Talbot JD, Evans AC, et al. Distributed processing of pain and vibration by the human brain. J Neurosci. 1994;14:4095-4108.

[123] Peters ML, Schmidt AJM. Psychophysiological responses to repeated acute pain stimulation in chronic low back pain patients. J Psychosom Res. 1991;35:59-74.

[124] Hart DL, Isernhagen SJ, Matheson LN. Witness: the truth please. Work. 1997;9:295-297.

DE Lechner, PT, is Associate Professor, Division of Physical Therapy, School of Health-Related Professions, The University of Alabama at Birmingham, Bishop 102, 900 19th St S, Birmingham, AL 3:5284 ([e-mail]), and President and founder of ErgoScience Inc, Birmingham, Ala. Address all correspondence to Ms Lechner.

SF Bradbury, MAOM, ATC, is President and thunder of Quality Essential Health Systems Inc and Co-Director of WorkHab Australia, Dallas, Tex.

LA Bradley, PhD, is Professor of Medicine, Division of Clinical Immunology and Rheumatology, The University of Alabama at Birmingham.