Pain, Impairment, and Disability in the AMA Guides Part 3

Friday, May 5, 2006 | 0

By James P. Robinson, Dennis C. Turk, and John D. Loeser Journal of Law, Medicine & Ethics, 32 (2004): 315-326.
Copyright 2004 by the American Society of Law, Medicine & Ethics.

Back injuries have a bad reputation. The workman looks upon them with apprehension, the insurance company with doubt, the medical examiner with suspicion, the lawyer with uncertainty... The medical examiner is faced with the difficulty of estimating the true value of the subjective symptoms in the comparative absence of physical signs. His suspicion is born of the frequent disparity between these two. "This prophetic statement made almost 100 years ago highlights an ongoing problem - how people who are incapacitated by painful conditions such as chronic low back pain can be evaluated consistently and equitably for purposes of disability compensation. We confronted this vexing problem when we participated in writing the chapter devoted to impairment associated with pain (Chapter 18) of the American Medical Association's Guides to the Evaluation of Permanent Impairment, 5th edition (AMA Guides 5th). Our purposes in this essay are to clarify several important conceptual issues associated with the assessment of pain among disability applicants, and to articulate reasons why pain should be considered in impairment and disability ratings.

I Clarification Of Concepts

As the above discussion of the treatment of pain in the AMA Guides 5th implies, any attempt to address this complex issue must start with concepts that are internally consistent, and with postulated relations among concepts that are testable. In particular, it is necessary to clarify the meanings of "impairment" and "disability", and the relations between the two.

As noted earlier, we conceptualize impairment as referring to derangement in the structure or function of organs or body parts, while disability refers to limitations in the behavior of the whole person that are imposed by medical conditions. Implicitly, the concept of disability assumes that healthy individuals are capable of performing a variety of voluntary acts if they choose to do so, and that disabled people experience limitations in these choices because of their medical conditions. There are many kinds of disability, since there are many arenas in which people's activities might be restricted because of a medical condition. For example, a person might be disabled in the sense of being unable to work, or disabled in the sense of being unable to carry out routine ADLs.

Society is interested in disability. Medical evaluations - including impairment evaluations - are useful to the extent that they provide valid medical data that help disability agencies correctly identify applicants who are truly disabled. More specifically, if an individual claims that she is unable to carry out certain acts because of a medical condition, a disability agency has the task of determining: (1) whether the alleged activity restrictions exist; (2) whether the applicant has a significant medical condition; and (3) whether the medical condition can plausibly be viewed as the cause of her activity restrictions. The agency awards disability benefits only if it concludes that all 3 of the above conditions hold. Medical evaluations are used to address items 2 and 3.

In principle, the validity of impairment evaluations of disability applicants can be determined by examining evidence about the causal role of impairment in disability. In order for impairment to be viewed as a cause of disability, a necessary condition is that there be a statistical association between impairment and disability. Thus, it appears that validity research on the AMA system should consist of prediction studies in which impairment is the independent variable and disability is the dependent variable (see below). Unfortunately, this seemingly simple statement about the appropriate methodology for validating the AMA system is greatly complicated by ambiguities in the meanings and methods of measuring "disability" and "impairment." Work disability is the type of disability most frequently sought by applicants who undergo medical evaluations for which the AMA Guides 5th is used. Thus, it might seem appropriate to validate the AMA Guides 5th by studying relations between impairment and work disability.

However, observers over at least the past 75 years have noted that the ability of an individual to work depends on a host of factors, only some of which are medical. The AMA Guides 5th recognizes this complexity, and makes it clear that impairment evaluations provide information about only the "medical component" of disability. It states: "Physicians have the education and training to evaluate a person's health status and determine the presence of absence of an impairment... The impairment evaluation, however, is only one aspect of disability determination. A disability determination also includes information about the individual's skills, education, job history, adaptability, age, and environmental requirements and modifications. "The widely held belief that work disability is a product of both medical and non-medical factors implies that statistical associations between results of medical evaluations and severity of work disability would be expected to be modest.

In contrast, there are reasons to anticipate that medical factors might be much more closely associated with disability in the sense of limitations in an individual's ability to perform ADLs. First, whereas the physical demands of different jobs vary substantially, all people face essentially the same physical requirements as they engage in ADLs such as dressing and toileting. Second, there is convincing evidence that economic incentives influence work behavior, presumably by influencing workers' motivation to continue employment. In contrast, it is plausible to assume that virtually all people are highly motivated to maintain independence in basic ADLs, and will do so unless their physical limitations give them no options. For both these reasons, it is plausible to anticipate that medical factors will be more strongly associated with ADL deficits than with work ability.

We believe that ADL deficits should serve as the validity criterion for medical assessments that are performed on disability applicants. We base this conclusion in part on the assumption that ADL deficits are likely to be closely related to medical factors. Also, we note that biomedical researchers routinely assume that ability to perform ADLs is an index of an individual's medical status, and use subjects' self-reported ability to perform ADLs as an outcome variable in clinical trials.

Just as the term "disability" has multiple meanings, "impairment" can also be measured and construed in multiple ways that bear on the methods that are appropriate to validate it. Sometimes impairment is used in a qualitative sense; namely, disability agencies require that an applicant has some minimal amount of impairment, and that the impairment typically causes limitations that are at least qualitatively similar to the ones alleged by the applicant.

The United States Social Security Administration employs a variant of this approach. It requires that disability applicants have impairments that are qualitatively congruent with their complaints, and broadly classifies applicants on the basis of the severity of their impairments into 3 groups: (1) those who are automatically eligible for benefits, (2) those whose applications are automatically rejected, and (3) those who are intermediate. Additional non-medical data are obtained on applicants in the third group to determine their eligibility for benefits.

In contrast, the AMA Guides 5th indicates that impairment can be quantified on a scale from 0% to 100%, and at least implies that the severity of an individual's impairment is predictive of the severity of his/her disability (in the sense of ADL deficits). Thus, for example, a person with 30% WPI because of a lumbar spine condition would be expected to have more ADL limitations than an individual with 5% lumbar spine impairment.

Moreover, as described above the AMA system uses the concept of whole person impairment. This construct permits comparisons to be made between people with entirely different medical problems that lead to qualitatively different patterns of ADL restrictions. For example, the AMA Guides 5th indicates that a person with renal disease and a creatinine clearance of 42-52 ml/minute has approximately the same amount of WPI as a person with a lumbar spine condition warranting DRE Category V impairment. We are not aware of any literature that validates the WPI construct.

We believe that research to validate impairment assessment systems should examine relations between quantitative indicators of impairment and deficits in the ability of individuals to perform ADLs. We believe that it is appropriate to examine these relations within a category of disorder (e.g., lumbar spine conditions) but that it is premature to attempt to make comparisons across conditions that have qualitatively different effects on the ability of people to function. In the discussion below, we assume that the ability of people to perform ADLs is the gold standard against which impairment evaluations should be validated.

II Arguments For Including Pain

With the above clarifications as a backdrop, we are now ready to address arguments for and against the inclusion of pain when the AMA system is used to evaluate disability applicants. Our view is that the validity of medical assessments of disability applicants who report persistent pain will be enhanced if the assessments include data about applicants' impairment (as described in the AMA Guides 5th) and also an assessment of their self-reports about their pain and its impact on their ability to function. Unfortunately, the scientific evidence to date is so meager that it does not provide convincing confirmation or disconfirmation of this thesis. However, we believe that several considerations support it:

1. The assertion that ADL deficits can be predicted accurately on the basis of indices of organ or body part dysfunction implies that measurable biological abnormalities exert an overwhelming influence on the behavior of people who have sustained injuries. This assumption is routinely contradicted by clinical encounters with disabled people, whose ADL deficits typically reflect a complex mixture of objectively measurable functional losses (e.g., weakness secondary to nerve injury or ankylosis of a joint) and a variety of subjective factors such as pain inhibition.

2. The assertion in point number 1 above also entails a "machine model" of behavior in which parts "break" or "mechanically fail" independent of pain or other subjective experiences of the injured person. This model is sometimes appropriate. For example, a person with a median neuropathy at the elbow might cease typing after a few minutes because his affected median nerve has ceased conducting impulses, so that he is unable to activate median-innervated muscles in his forearm or hand. However, it is also possible for the person to stop typing because of subjective factors such as progressive discomfort or fear of further injury, even though his median nerve is still capable of conducting impulses.

Among people with intact central nervous systems, mechanical failure and discomfort (or other subjective factors) as causes of activity limitations can, in principle, be distinguished operationally by means of electric stimulation of the relevant unit of the musculo-skeletal system. By a unit of the musculo-skeletal system, we mean a joint plus the nerves and muscles necessary for motion of the joint. Electric stimulation of appropriate motor nerves tests musculo-skeletal units in their entirety. Thus, for example, if a person demonstrates vigorous eversion and dorsiflexion of her ankle in response to electric stimulation of her peroneal nerve at the fibular head, an examiner would appropriately conclude that the entire unit consisting of the peroneal nerve, the peroneal muscles and tendons, and the ankle and subtalar joints was mechanically intact.

If this person reported that she was unable to voluntarily evert or dorsiflex her ankle, the alleged incapacitation would need to be attributed to subjective factors such as discomfort rather than to mechanical failure.

Mechanical failure often explains activity limitations in conditions where a critical component of a musculo-skeletal unit is completely non-functional (e.g., ankylosis of the wrist or a complete peroneal neuropathy at the fibular head). However, we maintain that in conditions that compromise function of a unit of the musculo-skeletal system but do not eliminate it completely, people essentially always terminate activity because of subjective discomfort or anticipation of pain that occurs before they reach the point of mechanical failure. In essence, we agree with the distinction between mechanical failure and "unbearable sensations" articulated by Osterweis et al.

Our hypothesis does not negate the possibility that indicators of organ or body part dysfunction can be used validly in the assessment of activity limitations. It is possible, for example, that in many conditions there is a strong correlation between these measures and the activity threshold at which people experience so much pain that they terminate an activity. In this instance, severity of organ or body part dysfunction could be viewed as a surrogate for the pain that acts as the proximate cause of activity limitations. However, if our hypothesis is correct, it would provide prime facie evidence for the importance of pain in the activity limitations that are central to the validity of impairment ratings. At the very least, it would place the burden of proof on a system that attempted to link indices of organ or body part derangement to ADL deficits without considering pain.

3. From an empirical standpoint, impairment rating can be conceptualized as a problem of prediction. An examiner gathers data about a claimant, then uses tables and formulas in the AMA Guides 5th to determine the claimant's percentage impairment. The impairment percentages are validated by empirical data or expert consensus about their correlations with ADL deficits.

The process can be modeled by a regression equation: Predicted Y = f(X1, X2,...Xn; x1, x2,...xm) Where: Predicted Y = a claimant's predicted ability to perform ADLs X1, X2,... Xn = the claimant's "scores" on indices of organ or body part derangement x1, x2,... xm = the claimant's "scores" for self-reported pain The empirical question is: Are indicators of organ or body part derangement sufficient to maximize the ability of the regression equation to predict disability, or does the predictive ability increase when these variables are supplemented by self-report data regarding pain? The answer to this question depends on correlations between: (1) organ or body part indices and self-report variables; (2) organ or body part indices and ADL ability; and (3) self-report variables and ADL ability. Correlations among these classes of variables will, of course, depend on the medical condition under consideration, and on the specific measures used to assess relevant variables. It is beyond the scope of this essay to review research in this complex area in any depth.

However, in the specific case of low back pain, self-report measures demonstrate the following statistical properties: (1) they are only modestly associated with objective indicators of spine dysfunction and (2) they are substantially correlated with functional outcome measures such as ability to perform ADLs. Thus, a system to predict functional status among low back pain claimants is likely to be more accurate if it incorporates self-report data in addition to indices of spine function than if it relies solely on indices of spine function.

4. An assessment system that relies exclusively on indices of organ or body part derangement prevents the inclusion of pain in an a priori fashion. We believe that exclusion of pain from impairment-disability evaluation systems before any serious effort has been undertaken to see whether it can be integrated into the systems represents premature conceptual closure.

III Arguments Against Including Pain

In a general way, it appears that some view a system for assessing pain among disability applicants as unnecessary, cumbersome, ambiguous, and as an aberration diverging from the objectivity manifested in the system(s) described in other chapters of the AMA Guides 5th. The primary concerns appear to include:

1. It is redundant to develop a system for assessing pain, because pain has already been factored into the ratings that are described throughout the AMA Guides as suggested in Chapter One.

This issue is crucial. As noted earlier, the AMA Guides 5th states that expert panels considered pain when they developed impairment ratings for various medical conditions.

But it does not discuss the process that panels followed as they considered pain, and glosses over many of the conceptual issues related to pain assessment that we raise in the present essay. Thus, we are skeptical that the panels considered pain in the detailed, systematic manner that we believe is appropriate. In order to make a convincing case for its statement in Chapter 1, the AMA Guides 5th would need to provide details about the decision-making of its panels, and to address the multiple ambiguities that complicate the assessment of pain.

2. The assessment of pain insinuates unscientific, subjective methods into a system that is otherwise objective and scientific.

This concern rests on questionable assumptions. One is that there are clear criteria by which objective and subjective findings can be distinguished. Although disability agencies and insurance companies routinely demand that medical examiners rely on objective findings when they perform impairment and disability examinations, we are not aware of any published discussion of the criteria by which objective and subjective findings can be distinguished. At least four criteria might be used to identify objective findings. They could be construed as findings that: (1) are not influenced by voluntary behavior of a patient; (2) are simple, in the sense that they do not require complex judgments by an examiner; (3) are not biased by prejudices of an examiner; (4) demonstrate adequate inter-rater reliability when different examiners elicit them.

We believe that the fourth criterion is the most important for establishing that a finding is objective, and that research might well show that this criterion is attainable even for findings that are not objective by the other three criteria. A major implication of the above is that the determination of whether or not a finding is objective needs to be made on the basis of empirical data rather than "armchair analysis" or deference to expert opinion. Examiners may or may not be able to achieve high levels of inter-rater reliability when they assess self-report data. This point is made clearly in research on psychiatric diagnosis. Although psychiatric diagnosis requires clinicians to make complex inferences about behavior over which patients potentially have voluntary control, a substantial body of research indicates that diagnostic assessment instruments such as the Structured Clinical Interview for DSM-IV have acceptable reliability and validity in the diagnosis of various disorders. Thus, research on psychiatric diagnosis indicates that trained observers are capable of making complex judgments based on patients' self-reports that meet the scientific requirements for reliability and validity. Research might demonstrate a similar result for the assessment of pain among disability applicants.

A second assumption is that the parts of the AMA system that do not involve the assessment of pain (e.g., the DRE category for assessing lumbar spine impairment) are scientifically established and objective. The DRE system might intuitively seem to be "more scientific" than a pain assessment system, but this has not been empirically demonstrated. In fact, there is essentially no scientific information about the reliability and validity of any of the procedures described in the AMA Guides 5th for rating musculo-skeletal or neurologic injuries. Moreover, the DRE system is not as objective as it might seem on casual inspection. In actual practice, examiners make multiple subjective judgments when they use the system. As one example, consider the results of manual muscle testing in the lower extremities of a person with a lumbar spine problem. The strength that a person demonstrates is dependent on the integrity of the spinal nerves that innervate his or her lower extremity muscles, but also on subjective factors such as his or her pain level and willingness to cooperate. If the person demonstrates any weakness, the examiner makes an inference about the sincerity of the effort, and decides whether to factor weakness into his or her impairment award, or to discount the weakness as not reflecting a "true" disorder.

3. If pain is considered during medical evaluations of disability applicants, applicants will be able to "game" the system by exaggerating the severity of their pain and the activity limitations imposed by it.

This concern embodies distrust of disability applicants, along with doubts about the ability of examiners to assess pain scientifically (concern 2 above). If pain cannot be assessed rigorously, exaggeration of pain by disability applicants may not be discernable to examiners, so that applicants who report severe pain may receive excessive disability awards.

Although we recognize the legitimacy of this concern, we believe that it represents a selective focus on the problems associated with evaluating pain, without attention to counterbalancing considerations. Basically, the issue of whether to include pain in the medical assessment of disability applicants involves a trade off between the relevance of various factors and the ease with which they can be measured. In many musculo-skeletal and neurologic injuries, objective indices of organ or body part derangement are relatively easy to measure, but have only modest relevance to disability. In contrast, pain is more difficult to measure, but, in our opinion, is often highly relevant to the limitations that disability applicants demonstrate. Thus, it is an oversimplification to reject pain assessment on the basis of measurement problems without simultaneously considering the issue of relevance. Empirical research rather than "armchair analysis" is needed to determine whether the inclusion of pain leads to a net improvement in the assessment of disability applicants and, more specifically, whether applicants are able to game the system by exaggerating their pain.

4. If examiners perform formal pain assessments on patients in the course of impairment evaluations, the impairment ratings that the patients receive will increase, thereby creating increased liability for the insurance industry. Moreover, regardless of the scientific merits of a pain assessment system, its introduction will create uncertainty for insurance carriers, so that it will be difficult for them to plan for the liabilities they will face.

Members of the business community and the insurance industry have raised these concerns. We are well aware of the possibility that disability policies can have significant financial consequences for businesses,44 but believe that the above concerns are misplaced for two reasons. First, insurance companies routinely face the challenge of revising their products in response to changes in their exposure to claims. For example, the insurance industry had to adjust to the potential of additional terrorist attacks after the events of September 11, 2001. Similarly, we believe they could adjust their premiums and fees if changes were made in the way disability benefits are awarded to people with painful conditions. Second, our central thesis is that the medical assessment of disability applicants can be improved by considering subjective information about claimants' pain in addition to objective information about dysfunction of their organs. If this turns out to be valid, we believe that the societal need to maximize the fairness of disability evaluation should take precedence over the desire of insurance companies to maintain the status quo.

IMPLICATIONS

A decision by the editors of the AMA Guides 5th to develop a systematic framework for assessing pain and to require a pain assessment for every individual who reports significant problems with pain would have multiple implications.

1. At the most basic level, it would formalize the idea that the subjective experiences of disability applicants - especially ones regarding pain - are important enough to warrant consideration during medical evaluations. Conversely, it would reject the conceptual model that medical evaluations can rely exclusively on objective data regarding the functioning of organs and body parts of applicants.

2. It would force editors of the Guides to address conceptual issues that we have raised above; namely, to reconsider the definitions of "impairment" and "disability," to consider differences between objective indicators and subjective factors as causes of disability, and to consider the methodologies that should be followed in validation research on the AMA system. These issues have been addressed cursorily and inconsistently in the AMA Guides 5th.

3. It would probably lead either to a re-definition of "impairment," or to an acknowledgment that the medical evaluation of a disability applicant involves more than an assessment of the applicant's impairment. That is, the Guides would have to address the implications of the fact that pain and other subjective factors can contribute to disability, but that they cannot be construed as derangements of organs or body parts.

4. Systematic inclusion of pain would probably act as a stimulus promoting research on the validity of the AMA evaluation system in painful conditions. Presumably, skeptics would demand such research. Thus, concerns about the inclusion of pain might lead to a reversal of the research vacuum that currently shrouds the AMA system.

5. In the process of systematically including pain in the AMA system, editors of the Guides would be forced to consider how examiners should combine information about pain with information about applicants' impairment. It is beyond the scope of this paper to discuss our views about techniques for incorporating pain in detail. We have done this in other publications. 45 But a few general points about combining data about pain with impairment data deserve brief mention. First, we believe that the assessment of pain should be performed by physicians and should be incorporated into the medical evaluation of disability applicants.

We recognize that it is a challenge for any professional to assess accurately the subjective factors that influence an applicant's disability, but we believe that physicians have two advantages over claims adjudicators or other professionals involved in evaluating disability applicants. One is that they have the expertise to compare patients' selfreports to their objective findings of organ or body part derangement. Also, physicians have abundant experience in assessing the subjective impact of injury or illness in clinical settings, and factoring these assessments into their treatment plans. For example, if a patient reports severe pain associated with a wound, the treating physician is likely to prescribe pain medication in addition to treatment designed to facilitate healing of the wound. We believe that physicians should also elicit and interpret self-report data when they perform medical evaluations on disability applicants.

Second, we believe that data about pain should be collected systematically, so that it is possible to determine how evaluators arrive at conclusions, and to identify points of agreement and disagreement when different examiners evaluate the same patient. The protocol described in Chapter 18 of the AMA Guides 5th is an example of a systematic approach to the evaluation of pain. Third, we believe that an examiner who assesses pain in a disability applicant has no choice but to consider the credibility of the claimant, and to factor an assessment of this credibility into his or her overall conclusions. Procedures for assessing credibility are described in Chapter 18. Fourth, as discussed above, we believe that it is fruitful to conceptualize disability evaluation as a problem of prediction, in which inferences about the an applicant's disability (i.e., limitations in ability to perform ADLs) are made on the basis of 2 broad classes of variables - indicators of impairment and information regarding pain and related subjective factors.

The relative weights given to objectively measurable impairment versus pain in determining disability will vary from one musculo-skeletal/neurologic condition to another.

6. Systematic inclusion of pain would almost certainly lead to changes in the procedures followed by expert panels of the Guides when they develop methods to rate specific conditions. Since pain is intimately intertwined with other manifestations of diseases and injuries, panels would almost certainly have to consider pain as they consider various disorders, rather than shunting pain to a separate chapter of the Guides. If this model is followed, experts in pain would need to interact with specialists in orthopedics, neurology, neurosurgery, and so forth. In the development of assessment tools and impairment schedules for these conditions. If this type of interaction occurs, it would probably be feasible to eliminate a chapter of the Guides devoted specifically to pain, since the effects of pain would be considered in relation to other effects of musculo-skeletal/ neurologic disorders on a condition-bycondition basis.

SUMMARY AND CONCLUSIONS

We have reviewed the basic concepts of impairment and disability and situated them within their historical social context. We outlined some of the difficulties in separating these concepts and described some of the ambiguities within the AMA Guides 5th. Finally we considered arguments for and against the inclusion of pain in impairment-disability evaluation systems. In our judgment, the considerations in favor of including pain are more persuasive than those against it.

Consideration of pain assessment among disability applicants has forced us to confront several basic issues regarding the assessment of impairment and disability, including the definition of objective factors, the relative contribution of objective and subjective factors to activity limitations, and the types of information that are appropriate for health care providers to consider when they evaluate disability applicants. We have also had to examine critically the entire logic of the AMA system. Along with others, we have found that the system addresses the fundamental concepts of impairment and disability inconsistently, and that the lack of empirical research on the system hampers rational discussion of whether or not it should be changed.

Our goal in this essay has been to describe how these problems affect the consideration of assessment of pain.

Many systems and concepts have been proposed to describe how medical factors and environmental or social factors contribute to incapacitation. For example, Nagi outlined a system that used four constructs: active pathology, impairment, functional limitation, and disability. In 1980, the World Health Organization described a system based on the constructs of pathology, impairment, disability, and handicap. In a revised system, the WHO identified health conditions, impairments, activities, participation, and contextual factors as basic constructs. The plethora of systems and concepts is an index of the subtlety of relations among the factors that cause individuals to become incapacitated. In the face of this complexity, we believe that a fundamental distinction needs to be made between analyses at the level of organs or body parts and analyses at the level of the whole person. We therefore propose that impairment be consistently defined in terms of organ or body part derangement, and disability in terms of activity limitations at the level of the whole person. We also propose a distinction between objective factors and subjective factors that may influence an injured person's ability to perform ADLs following injury. We believe that when health care providers examine disability applicants, they should evaluate both objective measures of impairment and subjective reports of applicants - especially ones regarding pain.

Finally, we are concerned that although millions of disability evaluations are performed each year and the prevalence of pain is high among disability applicants, the scientific literature on how to address the needs of these applicants is virtually non-existent. Our hope is that the present essay will stimulate debate on this important subject.

ACKNOWLEDGEMENTS

Preparation of this manuscript was supported in part by grants from the National Institute of Arthritis and Musculo-skeletal and Skin Diseases (AR/AI44724, AR47298) and the National Institute of Child Health and Human Development/National Center for Medical Rehabilitation Research (HD33989) awarded to the second author

-------------------

This article is republished with the permission of The American Society of Law, Medicine & Ethics 765 Commonwealth Avenue Suite 1634 Boston, Massachusetts 02215 Tel: 617.262.4990 - http://www.aslme.org/

The views and opinions expressed by the author are not necessarily those of workcompcentral.com, its editors or management.