Recommendations to Guide Revision of the Guides to the Evaluation of Permanent Impairment

The American Medical Association's (AMA's) Guides to the Evaluation of Permanent Impairment, Fourth Edition¹ is the most commonly used tool for rating permanent impairments in the United States. Physicians in all specialties use it to assist in the evaluation of patients for disability programs. The purpose of this article is to encourage serious revision of the Guides and to make 9 specific recommendations that may assist the AMA in the revision process.

Originally published as a series of articles in JAMA, the Guides has been revised periodically since its initial publication as a single volume in 1971.^{2 - 5} In mid-1997, the AMA convened a 7-member steering committee to advise on the development of a fifth edition. That committee (on which all authors of this article served) was disbanded in June 1998. Subsequently, spokespersons for the AMA have indicated that major revision of the Guides is no longer planned for the next edition.⁶

The Guides is a tool to convert medical information about permanent impairments into numerical values. Each chapter focuses on a single organ system and provides a description of the diagnostic and evaluative methods for assessing specified impairments. Each impairment is assigned a rating, expressed as a percentage of loss of function for that system. Organ-based ratings are then translated into impairment ratings for the whole person, termed whole person impairment (WPI). For example, amputation of the index finger of either hand is considered a 20% impairment of the whole hand, an 18% impairment of the upper extremity, and an 11% WPI.¹ (pp18-20) Finally, the Guides combines multiple WPIs into a single rating by using the formula [A + B(1 − A)], where A is the rating for the first impairment and B is the rating for the subsequent impairment, thus creating an asymptotic curve toward 100%.¹ (p322)

The Guides is used for evaluation and adjudication of disability benefits in more than 40 state workers' compensation programs because it provides a standardized mechanism to generate ratings that quantify WPI.^{7 - 10} Disability programs often need to assign a numerical value for the extent to which an applicant is impaired by a physical or mental condition. The Guides provides an apparently attractive mechanism for arriving at this value: there is a presumption that medical ratings are scientific, unbiased, and reliable; it is drafted by recognized experts; and it is published by a respected medical organization.

Nevertheless, 2 types of serious and legitimate criticisms have been directed at the Guides. First, criticisms focus on internal deficiencies, including that it fails to provide a comprehensive, valid, reliable, unbiased, and evidence-based system for the rating of impairments^{11 - 19} ; that the impairment ratings do not reflect perceived and actual loss of function and quality of life^{16 - 17} ; and that the numerical ratings represent legal fiction, not medical reality.^{11 ,14 - 15 ,20 - 22} California, the largest state workers' compensation system, has declined to use the Guides largely due to these concerns.

The second set of criticisms focuses on how workers' compensation systems use the Guides' ratings. The concern is that the ratings are improperly used as a substitute for a full assessment of the impact of impairment on work and nonwork capabilities, and that therefore workers receive inappropriate compensation. Although the criticisms regarding the use (and misuse) of these ratings may have merit, this article does not address this issue.

Because there is a need for a valid and reliable system for rating impairment, the Guides plays an essential role. The following 9 recommendations are intended to provide a framework to address concerns regarding the internal deficiencies of the Guides. We understand the difficulty of designing a perfect impairment rating system, and we do not attempt to solve all the difficult issues here. Our hope is that serious public discussion of the application of these recommendations will yield a substantially improved Guides.

The Guides states that it provides the tools for assessing pathology and impairment (at the organ system level) and functional impairment of the whole person (the impact of the organ system loss on the ability of the individual to perform discrete functions, tasks, or activities of daily living). As the current edition notes, assessments of impairment generally involve medical expertise and information that is observable and ascertainable in the medical setting.¹ (pp1,315-317)

The Guides explicitly states that it is not intended to rate disability¹ (pp2,5) and notes that impairment and disability are not synonymous.¹ (pp1,2,317) Evaluation of disability requires nonmedical judgments that are generally outside the scope of physicians' expertise.^{23 - 30} Disability, according to the definition adopted by the Institute of Medicine and others, involves limitations in the ability of an individual to perform complex occupational, social, and family roles and therefore focuses on the interaction between an individual's own functional status and environmental and socioeconomic factors.^{26 - 28} For purposes of clarity, we note that this definition differs somewhat from that currently used by the World Health Organization³¹ and the more expansive definition in the Americans with Disabilities Act.³² Unfortunately, there is not space here to explore these differences.

Despite its stated intent, the Guides does not maintain a clear focus on functional impairment. First, it blurs the line between impairment and disability by including disability-related roles (such as "occupation" and "social and recreational activities") in the lists of activities relevant to impairment.¹ (ppvi,317) As a result, the Guides defines impairment so broadly as to include disability, effectively confusing the 2 concepts, and thereby encouraging inappropriate use of the ratings as a measurement of disability.¹⁹

Second, despite the introductory assertion that the impairment ratings reflect functional loss and capacity, this concept is not applied consistently. For example, the spine section of the musculoskeletal chapter focuses on structure rather than function, using a "diagnosis-related estimate" or "injury model."¹ (pp100-111) Under this model for estimating impairment, physicians are directed to ignore both "developmental" changes of the spine (including osteoarthritis), irrespective of etiology, and the postinjury sequelae and effects of treatment (including surgery), and thus do not include a full evaluation of the permanent impairment.¹ (p100)

Third, the Guides provides inadequate analyses of key components for rating functional loss, including what activities, functions, or tasks should be included in assessing functional limitations; the relationship between the numerical ratings and the loss of ability to perform these activities; the relative importance of particular activities; and the relationship between functional inability to perform specific tasks or activities and assessment of WPI. The selection, valuing, and measurement of functional limitation in performance of activities and tasks, a core issue in the development of an evaluative system for impairments, should be fully explained in the Guides. It would be advisable for future editions to draw on other functional limitation research.^{26 - 27 ,33} The Guides should identify explicitly the underlying choices being made with regard to selected activities and apply these choices consistently across organ systems.

The Guides' numerical ratings for organ system impairment and WPI are based on consensus opinions of chapter authors about the severity of particular conditions and have changed little over the years. Serious questions have been raised about the ratings' validity.^{11 - 22} With rare exception, recent research has focused on the reliability and reproducibility, not the validity, of the ratings.^{34 - 37}

While the authors of earlier editions of the Guides may have had no alternative but to rely on consensus to produce the ratings, there is now a rapidly expanding literature concerning the measurement of functional limitations that invites the development of an evidence-based impairment rating system. A number of possible approaches to validation could assist in the development of a system to convert physiological measurements to impairment ratings.^{26 ,33 ,38 - 39} Validation of the rating system requires that the scale for the ratings be defensible, that the ratings be both comprehensive and consistent, and that the system be externally validated. Every attempt must be made to relate each impairment rating to an overall understanding of its impact on the ability of the individual to perform selected tasks and activities. External or "construct" validation also can be based on available data regarding the level of loss that people associate with particular medical impairments. Population-based health status and medical outcome studies would be useful in this process. Evaluative techniques to measure specific limitations of function also should be considered.⁴⁰

The scientific basis for all ratings should be included, as was done by the Agency for Healthcare Research and Quality (formerly AHCPR) in its recommendations for guidelines for the treatment of low back pain.⁴¹ The Guides should provide a thorough summary of the scientific evidence, including its strengths and limitations, for each section. If ratings must be consensus-based because of the lack of valid data, the Guides should explain with specificity the basis on which the ratings are derived. Biases should be removed, or, at a minimum, specifically discussed. Normative judgments that are not data driven should be made explicit.

Finally, there must be clear boundaries between scientific and medical issues and questions that are of an economic or policy nature. At no time should the Guides' rating system disregard functional limitations because of concern that the ratings will generate excessive cost in a social insurance system that chooses to use the Guides. Economic implications of the ratings can be addressed effectively in the legislative, administrative, and political arenas.

The scale used to generate permanent impairment ratings is a core component of the validity of the Guides. Assuming the use of a 100-point scale, 2 critical questions must be answered: What is the appropriate definition of 0% impairment? What is the appropriate definition of 100% impairment? Points along the scale can be then be estimated.

The top of an impairment scale (100%) should reflect a level of functional loss related to inability to perform the specific tasks that are necessary for independent daily life. Instead, the Guides defines 95% to 100% impairment as "a state that is approaching death."¹ (p8) The high level of impairment required for a 100% WPI rating results in the devaluation of significant physical and mental impairments. For example, in the fourth edition of the Guides, a mental status impairment that "requires directed care under continued supervision and confinement in home or other facility" is rated at 30% to 49% WPI¹ (p142); severe paroxysmal disorder of such frequency that it limits activities to those that are supervised, protected, or restricted is also rated at 30% to 49% WPI¹ (p143); a patient capable of spontaneous respiration but to such a limited degree that he/she is confined to bed can be rated as low as 50% WPI.¹ (p149)

The setting of an excessively high standard for 100% impairment essentially depresses the entire scale so that individuals' functional losses are not accurately reflected in their numerical ratings. The practical result of this is 2-fold: the Guides consistently underrates impairments as they are perceived by patients and independent observers,^{16 - 17} and benefit systems that place great weight on impairment ratings value some impairments at inappropriately low levels.

Furthermore, the current scale leads to the adoption of the asymptotic formula for combining impairments.¹ (p322) Given the definition of 100% WPI, clearly no one can be more than 100% impaired. Therefore, the Guides requires that each subsequent impairment be reduced in value. Thus, loss of a foot is valued at 25% WPI and loss of the second foot results in a total rating of 44% WPI.¹ (pp83,322) In reality, the combining of impairments in an individual can result in additive, less than additive, or greater than additive levels of functional loss. The current formula for rating multiple impairments always results in a less than additive result, an outcome that produces mathematical consistency but not accuracy.

The bottom of the impairment scale should reflect an unimpaired ability to perform the tasks and activities that have been specified as necessary to an assessment of functional loss. In the current Guides, 0% impairment is not defined. Consequently, different chapters of the Guides adopt different approaches to what constitutes a measurable (>0%) impairment. The implications of this are particularly striking in situations in which an individual is advised to avoid certain activities on a prophylactic basis.

In addition, although the Guides is intended to rate functional loss, it does not include any explicit discussion of the appropriate baselines for normal function, a critical component for a defensible rating system. Any discussion of normal values must address a variety of possibilities: Should baselines include known population variants such as age, sex, or race? Should baselines include adjustments for the fact that individuals who work are healthier than the general population? When individual baselines are known, such as when individuals are tested at the start of employment, should these influence an assessment of functional loss?⁴² Should baselines be derived from estimates of loss of preinjury capacity (in the absence of individual baseline measurements)? The Guides should explicitly define appropriate baselines and apply them consistently.

To the extent feasible, the rating system should produce consistent ratings when the same person is rated by different physicians, or when different people with equivalent impairments are rated by 1 or more physicians. The popularity of the Guides rests, in large part, on its claims of reliability. The current reliability of the ratings system has, however, been subjected to only limited testing and has not been demonstrated.^{36 - 37} To improve and validate the reliability of the Guides as a system for rating permanent impairment, the AMA should develop mechanisms for testing and reviewing the interrater and intrarater reliability of the Guides when used by practicing physicians.

A permanent impairment rating system should include all impairments, regardless of whether they may result in work disability. All conditions resulting in functional loss should receive a rating greater than 0% impairment.

The current edition of the Guides does not meet this test of comprehensiveness. For example, degenerative and terminal illnesses are not adequately addressed. Aggravation of a condition is predicated on the existence of a prior rated impairment and requires an increase in impairment of more than 3%.¹ (p316) Chronic headaches are not rated.

The Guides relies on objective tests of impairment, drawn from laboratory or diagnostic results, because they tend to yield reliable, reproducible results.³⁶ The current musculoskeletal chapter explicitly rejects, for example, functional tests influenced by subjective factors.¹ (p64) Although reliance on objective testing may seem reasonable and enhance the reliability of the Guides, the result is the exclusion of a large number of impairing conditions because the evaluative method is viewed as inadequately objective. In consequence, the Guides' comprehensiveness is reduced, threatening the validity of the rating system as a whole.

There is inevitable tension between the goals of comprehensiveness and validity and the goals of reliability and reproducibility. To the extent that subjectivity is introduced into the assessment of impairments, the reproducibility of results may suffer. However, an inclusive guide to impairment evaluation cannot ignore impairments that cannot be verified with currently available objective tests. To the extent that a balance must be achieved, the current Guides leans too far toward the requirement that rated impairments be justified by objective findings.

The assessment of pain presents special problems. Although subjective, pain is generally recognized to be an important cause of functional loss and disability.²⁷ Currently, pain is treated inconsistently. The Guides should develop a consistent approach to pain for all organ systems that standardizes the description and assessment of different types of pain and develops methods to test the effects of pain on a person's ability to perform activities.

Some gaps in comprehensiveness may be inevitable. To ensure that impairing conditions that are not included do not automatically receive a 0% impairment rating, the Guides should provide instructions to clinicians to rate such impairments and guidance regarding the underlying rating system.

The Guides should strive to achieve ratings that are consistent within chapters and from one organ system to another, based on an equivalent scale that focuses on the individual's loss of ability to perform selected activities. Unfortunately, the current edition has many internal inconsistencies. First, the methods for rating impairments vary significantly for different types of injuries and conditions. For example, ratings for some musculoskeletal disorders are derived from pretreatment diagnoses, not functional status. For other organ systems, impairment is rated after treatment is completed.¹ (p185)

Second, impairment ratings in different chapters do not show a consistent relationship to functional loss or capacity. For example, someone with coronary heart disease who has no symptoms while performing moderately heavy physical exertion may be rated at 29% WPI; conversely, someone with significant lower extremity impairment with atrophy, loss of reflexes, and numbness can receive a maximum WPI rating of 25%.¹ (pp102,178)

Third, there is no consistency in the ratings of impairments that involve future risk or restricted capacity. On medical examination the individual may exhibit no apparent functional loss. However, when confronted with a situation in which the underlying pathology is challenged, the individual's capacity to function may be reduced. This individual should receive a WPI rating greater than 0%. Compare the following approaches to this problem:

A patient with pain and swelling in the hand and wrist who performs repetitive motion activities at work has no remarkable findings from physical examination and electrodiagnostic studies; symptoms are noticeable only at work. The Guides notes, "It was clear the woman should not return to her former job." Rating: 0% WPI.¹ (p19)

The chapter on respiratory diseases notes, "Asthma presents a difficult problem in impairment evaluation because results of pulmonary function studies may be normal or near normal between attacks. Impairment estimate should be left to the physician's judgment."¹ (p164)

A 50-year-old man following myocardial infarction has normal findings on physical examination and chest radiography; the electrocardiogram shows Q waves in leads 2, 3, and F and flat T waves in the same leads. When exercising, his heart rate is 152/min, he has an adequate elevation in blood pressure, and no electrocardiogram pattern changes. Rating: 20% WPI; if uncomplicated recovery from an anterior wall infarction, rating would be 29% WPI.¹ (p177)

To achieve both reliability and validity, the editors of the Guides must develop a consistent approach to rating and direct the authors of the various chapters to follow this approach.

All interested parties, including treating physicians, should be able to use an impairment rating system based on instructions contained in the schedule. It should not be necessary for a user to seek other sources, or to rely on unstated premises or knowledge, to understand and apply the rating system.

The methods for rating impairments should strive to be relatively inexpensive, easy to teach, and easy to apply. When choices among evaluative techniques must be made, the chosen method should involve, whenever possible, equipment or tests that are readily available to the average treating physician. The ratings system itself should not require expensive or dangerous diagnostic procedures that are not medically necessary.

The numerical ratings for impairments must be accepted by the many participants in the workers' compensation and other disability programs, including treating physicians, employers, unions, administrators, and courts. Acceptability depends in part on the origins of the relative values and in particular on whether there is some scientific basis for the ratings.

The Guides has come under justifiable criticism for setting arbitrary ratings for conditions. The methods used to develop the numerical ratings of impairment must be improved or the credibility of the Guides will be jeopardized. Future editions need to mirror the scientific validation processes that guide the development and dissemination of new therapies and diagnostic tests. When evaluation of impairment must be based on normative judgments, the development of these judgments must be subject to open discussion and review.