NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) and National Research Council (US) Committee to Review the Social Security Administration's Disability Decision Process Research; Wunderlich GS, Rice DP, Amado NL, editors. The Dynamics of Disability: Measuring and Monitoring Disability for Social Security Programs. Washington (DC): National Academies Press (US); 2002.

Cover of The Dynamics of Disability

The Dynamics of Disability: Measuring and Monitoring Disability for Social Security Programs.

Show details

SSA's Disability Determination of Mental Impairments: A Review Toward an Agenda for Research

March 1999

(Updated October 2001)

Cille Kennedy, Ph.D.1

The Social Security Administration (SSA) operates two disability benefit programs; Social Security Disability Insurance (SSDI) for disabled workers and Supplemental Security Income (SSI) for disabled impoverished adults and children. Both of these programs come under periodic scrutiny. Of present concern is the process by which claims for disability benefits are adjudicated. In an effort to provide policymakers with a scientific base for future deliberations and indicated directions, the Institute of Medicine (IOM) and the Committee on National Statistics (CNSTAT) have been asked to examine the reliability, validity, adequacy, and appropriateness of SSA's current and proposed research activities as they related to the proposed redesign of the disability determination (Wunderlich and Kalsbeek, 1997).

The focus of this paper, commissioned by the Committee to Review the Social Security Administrations' Disability Decision Process Research (the committee), is on the determination of disability status of initial claims, based on mental disorders, for SSDI and SSI disability benefits. The scope of this paper covers the initial determination and emphasizes the medical aspects of the process. It is limited to the adjudication of adult claims. The paper draws heavily on the evaluation—contracted by the SSA and conducted by the American Psychiatric Association (APA)—of SSA's standards and guidelines used in the determination of claims based on mental disorders. Building upon this base, the paper reviews the conceptual model and taxonomy of the World Health Organization's (WHO) International Classification of Functioning, Disability and Health (ICF) (WHO, 2001) and related WHO disability assessment instruments, and their potential utility in the redesign of the determination process, and toward a cohesive agenda for research. The paper is intended to stimulate an agenda for research to inform future modifications of the disability determination whether or not a formal redesign is undertaken.


Statutory Definition of Disability

The foundation of the SSA's two disability programs is the statutory definition of disability in the Social Security Act. The same definition applies to both the SSDI and the SSI programs and is the standard that the SSA puts into operation for the determination of claims for disability benefits. The definition can be changed only by an act of Congress. According to the Social Security Act, the definition of disability is “Inability to engage in any substantial gainful activity by reason of any medically determinable physical or mental impairment which can be expected to result in death or which has lasted or can be expected to last for a continuous period on not less than 12 months” (Section 223(d)(1)(A)).

The term “substantial gainful activity” means work that is remunerated at a rate specified by regulations. As of January 1, 2001, the rate is $740 per month. In other words, an individual may not be earning more than $740 per month in order to be eligible for disability benefits. The statute further states that the physical or mental impairments must be so severe that the claimant cannot do any work in the national economy that exists in substantial numbers. It does not matter whether or not jobs are available in the local region or whether or not the person would actually be hired if a job existed. If a type of job exists in substantial numbers somewhere in the country that the claimant could do, then she or he is not given disability benefits.

Conversely, consideration is given to the person's age, educational level, and past work experience. A person nearing retirement age is treated differently by SSA than a younger, working age individual. The older person is more likely to be considered favorably for disability benefits. A person with a grade school education is not expected to be able to work at an available occupation that requires advanced educational expertise, and a person with a work history of manual labor is not expected to obtain an available position as a business executive. The bottom line for SSA disability is that the person cannot do the simplest, least demanding, existing work whether it is available or not.

The reason that an individual is unable to work must be due to a physical or mental impairment that is medically detectable. The statute goes on to define physical and mental impairments as resulting from “anatomical, physiological, or psychological abnormalities which are demonstrable by medically acceptable clinical and laboratory techniques” (Section 223(d)(3)). The effect of any one impairment or any combination of impairments is to be evaluated for severity in determining disability for work. If one impairment is so severe that the person cannot work, then the person is considered disabled. If a combination of impairments precludes work, then the person would be considered disabled even if no single impairment would be considered severe by itself (Section 223(d)(2)(C)).

Sequential Evaluation

The process by which SSA adjudicates initial claims for both SSDI and SSI disability benefits is called sequential evaluation. There are five steps in the initial process. Responsibility for completing these steps is divided between federal SSA workers in local SSA district offices (DOs) and state employees working in state Disability Determination Services (DDSs). The SSA contracts with state agencies, such as departments of social services or rehabilitation, to act as DDSs.

The first step in the sequential evaluation takes place at the local DO. The person claiming disability (or a representative) appears and applies for benefits. Here, the DO staff determines whether or not the individual is currently working according to the criteria established by the regulated amount considered to be “substantial” (currently $740 per month). If the person is earning at or above this level, the claim is denied at this step.

If the claim is not denied, then it is necessary to decide whether the claim should be processed for SSDI or SSI benefits, or both. Although the process is the same, the administrative criteria and cash award amounts differ. SSDI is an entitlement program for workers. There are specific work history requirements for SSDI. Workers pay into the Social Security system and therefore have the right to receive cash benefits if disabled. The SSI program is for people whose income and resources are below a certain monthly level and who are blind, aged, or disabled. Children are eligible for SSI. It is possible to receive both SSDI and SSI benefits simultaneously. Once the decision is made about which disability program the person is eligible for, the SSA DO staff requests that the applicant supply the necessary medical evidence and work history to support the claim of disability. The SSA DO staff then forwards the claim to the state DDS.

The second step begins when the application is received in the state DDS. Here a team collects and evaluates the medical evidence and work history. There must be sufficient medical evidence to substantiate a determination of the claimant's disability status. The team consists of a Disability Analyst and a Reviewing Medical Consultant. For claims based on mental impairments, the Reviewing Medical Consultant is usually a psychiatrist or clinical psychologist.

The Listings of Mental Impairments (the Listings) are the standard for the evaluation of the medical evidence for demonstrable signs, symptoms, and restrictions in daily life (described below). They are the translation of the medical component of SSA's definition of disability. The review of claims based on mental impairment is put into operation against the standards of the Listings using the Psychiatric Review Technique Form (PRTF) (described below.) The PRTF guides the decision-making process, conforming to SSA regulations (SSA Regulations, 404.1520 and 404.1520a) as to how decisions are to be made.

The Reviewing Medical Consultant determines whether an impairment exists, and—if so—whether the impairment is considered severe. An impairment for SSA is analogous to a diagnosis of one of nine alcohol, drug, or mental disorders. Severity is concluded on the basis of whether the claimant's condition results in slight or marked restriction of activities. If the impairment is found to be slight or ‘not severe,' the claim is denied on the basis of this medical consideration alone. If the impairment is considered severe, the claim continues in the sequential evaluation process.

The third step involves severe cases only. The assessment at this step inquires whether the impairments are so severe as to preclude work on the basis of medical evidence alone. In this step, the Reviewing Medical Consultant decides whether or not the claimant's impairment(s) “meet or equal” standards set by the Listings of Mental Impairments.2 If a case meets or equals the Listings as indicated on the PRTF, then the claimant is allowed benefits by this medical evaluation. If this severe case does not either meet or equal the Listings, it continues in the sequential evaluation.

Steps 2 and 3 are the only steps that permit a disability decision based on medical assessment alone. In step 2 the Reviewing Medical Consultant may determine that a claim is “not severe” and the claim is denied. The Listings of Mental Impairment are constructed in such a way that an individual who ‘meets or equals' them in step 3 of sequential evaluation cannot reasonably be expected to work and the claimant is awarded SSA disability benefits.

The fourth step applies to those severe claims that have not been found so severe as to be allowed disability benefits on the basis of medical evidence alone. For these claims, the nonmedical factors of age, education, and work history are taken into account. Another difference in this step is that the Reviewing Medical Consultant provides additional input into the decision by completing the Mental Residual Functional Capacity Assessment (MRFCA) (see description below); it is the Disability Analyst who combines the narrative summary of the MRFCA with the age, educational level, and work history of the claimant to determine whether or not the claimant is capable of working at the level of her or his past employment. This decision is made in light of jobs available in the national economy. If the Disability Analyst finds that the claimant can do previous work, the claim is denied. If the finding is that the claimant cannot do previous work, the claim continues one final step in the initial review.

The fifth step applies the same claim material to the question of whether the claimant can do any job in the national economy. If the Disability Analyst determines that the claimant can do work—irrespective of whether it is locally available or whether the person would actually be hired—then the claim is denied. If the person cannot do any work in the national economy, then the claimant is awarded disability benefits.

The next section describes the standards and guidelines upon which these medical judgments are based.

Listings of Mental Impairments

The Listings of Impairments (SSA, 2001) are published by the SSA and updated periodically. Chapter 12 of the Listings is devoted to mental disorders, otherwise known as the Listings of Mental Impairments. Major revisions to the Listings of Mental Impairments, currently being applied to claims for disability benefits, were published in 19853 and have since undergone relatively minor modifications. The 1985 revision was intended to bring the Listings in line with then-current psychiatric practice to reflect the APA's third edition of its Diagnostic and Statistical Manual of Mental Disorders (DSM-III) (APA, 1980). The process of this revision was unique in SSA's history. It was the first time that the SSA had sought outside expertise to revise its Listings. The APA, the American Psychological Association, and other mental health experts participated with the SSA in the process.

There are currently nine Listings of Mental Impairment designed to reflect the major psychiatric diagnostic categories likely to cause disability for work.4 There are three sets of criteria within the Listings of Mental Impairments: A, B, and, for a subset of categories, C.

The purpose of the A criteria for mental disorders is to medically substantiate the presence of a mental disorder. The categories for adults follow:

  • Organic Mental Disorders;
  • Schizophrenic, Paranoid and Other Psychotic Disorders;
  • Affective Disorders;
  • Mental Retardation;
  • Anxiety-Related Disorders;
  • Somatoform Disorders;
  • Personality Disorders;
  • Substance Addiction Disorders; and
  • Autistic Disorder and Other Pervasive Developmental Disorders.

The categories contain either two or three sets of criteria.5 The A criteria are essentially diagnostic-like symptoms. They are not exact replicas of the DSM-III but are analogous to them. Each of the categories, except mental retardation and substance addiction, lists clinical findings. With the noted two exceptions, the threshold for the A criteria is that at least one of the clinical findings must be present. For example, to fulfill the A criteria for the category of Schizophrenia, Paranoid and Other Psychotic Disorders, there must be medically documented evidence of persistence, either continuous or intermittent, of at least one of the following:


delusions or hallucinations;


catatonic or other grossly disorganized behavior; or


incoherence, loosening of associations, illogical thinking, or poverty of content of speech associated with one of the following:

  • blunt affect,
  • flat affect, or
  • inappropriate affect; or

emotional withdrawal and/or isolation.

The categories of Mental Retardation and Substance Addiction Disorders each differ and are not discussed as part of this review. They require special expertise and attention to detail beyond the scope of this paper.

The B criteria are applied to the category for which the A criteria are fulfilled. The purpose of the B criteria for mental disorders is to describe the functional restrictions that are incompatible with work and are associated with the mental impairments of the A criteria. The B criteria follow:

  • marked restriction of activities of daily living;
  • marked difficulties in maintaining social functioning;
  • marked difficulties in maintaining concentration, persistence, or pace; and
  • repeated episodes of decompensation, each of extended duration.

As an example, the SSA describes activities of daily living to include cleaning, shopping, cooking, taking public transportation, caring for one's grooming and hygiene, among others. These activities are referred to as activities of daily living and instrumental activities of daily living in the professional literature.

The C criteria apply to all categories except Mental Retardation, Personality Disorder, Substance Addiction, and Autistic Disorder.6 The C criteria are additional considerations for cases that do not reach the threshold of the B criteria. For example, the C criteria for Schizophrenia, Paranoid, and Other Psychotic Disorders are intended to compensate for such instances when claimants are living in supportive residential settings or are otherwise adapted to a special environment that could not be sustained if the person went to work. These C criteria also consider individuals who have a history of serious episodes of disorder or disability and are currently functioning at a relatively high level through the benefits of medication but whose delicate functional status would be jeopardized by the additional stress of work. In other words, the relatively high degree of functioning is attributed to the compensatory medications or supports. Work would jeopardize this accomplishment and would vitiate the level of functioning attained. The C criteria for Anxiety-Related Disorders are designed to accommodate individuals with agoraphobia who are totally unable to function outside their homes but can function successfully within the home.

The following description of the forms show how decisions are made that put the sequential evaluation into effect for the medical component of the disability determination.

Forms Used by the Reviewing Medical Consultant

The Reviewing Medical Consultant uses two forms in the sequential evaluation. The forms are used to document the existence of the medical condition and its impact on the domains of functioning related to the ability to work. The Psychiatric Review Technique Form puts the Listings of Mental Impairment into operation. The Mental Residual Functional Capacity Assessment is used to assess remaining functioning for claimants who are considered severe, but not sufficiently severe to be awarded benefits on the sole basis of the medical evidence using the PRTF. The MRFCA is intended to document what the person can do in spite of severe impairment. Unlike the B criteria of the Listings that ask the degree of limitation, the MRFCA intends to look at residual functioning—what the person can still do.

Psychiatric Review Technique Form This is designed to facilitate a review of the medical evidence and guide a medical decision as to the disability status of the claimant. The cover sheet, Section I, contains the summary of the medical review in two parts: the medical disposition and the category on which the medical disposition is based. The second page, Section II, provides space for the Reviewing Medical Consultant's notes. Following this is Section III, which lists the different categories of mental impairments along with their A criteria. At the top of each category are two checkboxes in which to indicate whether or not evidence of a cluster or syndrome exists that fits the particular diagnostic-like category. Beneath those two checkboxes, each of the category's A criteria is preceded by three checkboxes in which to indicate whether the specific item is present or absent, or whether insufficient evidence is provided in the medical information. The Reviewing Medical Consultant selects the one diagnostic-like category under which the claim will be reviewed and fills out the boxes for those items. If the A criteria are fulfilled, the Reviewing Medical Consultant then proceeds to the section that contains the four B criteria.

The page dedicated to the B criteria has two sections. The first is a chart that has the four B criteria, the areas of functional limitation listed on the left, and to the right, four or five checkboxes with which to rate the degree of limitation. Although degree of limitation is generally conceptualized as a continuum, for programmatic practicality, five intervals are identified. For example, restrictions in activities of daily living (the B1 criterion) can range from none, to slight, moderate, marked, and extreme.

These degrees of limitation are used to make two decisions: whether the claimant is (1) slightly limited (a step 2 denial) or (2) so severely limited that a benefit can be awarded (step 3). A slight impairment exists if all four B criteria are checked in the two left-hand columns (none, slight, never, or seldom). All other claims are considered severe. For a claim to be so severe as to meet or equal the Listings of Mental Impairments, two of the B criteria must be checked in the two columns to the right (marked, extreme, frequent, constant, repeated, or continual). As noted above, somatoform and personality disorders require that three of the B criteria must be so designated.

The C criteria are assessed as to their presence or absence.

This detailed description of the forms is not presented without reason. The judgments made by using the checkboxes result in two medically based disability determinations: denials for nonsevere claims and allowances for claims that are medically considered so severe as to preclude work. A “marked” limitation is considered a clinical decision. SSA describes “marked” as between a moderate and an extreme limitation. How sound are these decisions?

Mental Residual Functional Capacity Assessment This is used only for claims that are severe but have not been allowed disability benefits in the previous step of sequential evaluation, either meeting or equaling the listings. The MRFCA provides additional medical review for the Disability Analyst to combine with the nonmedical factors of age, education, and work history. The MRFCA is a checklist of 20 items subaggregated into four categories: (1) understanding and memory; (2) sustained concentration and persistence; (3) social interaction; and (4) adaptation. The form calls for a rating of limitation in the context of the individual's capacity to sustain the listed activity over a normal workday and workweek, on an ongoing basis. These items are rated on a three-point scale from “not significantly limited” to “markedly limited.” Two other checkboxes permit ratings of “no evidence of limitation” and “not ratable on available evidence.” “Not ratable” is to be used if the Reviewing Medical Consultant feels that there may be a limitation but cannot support a finding on the existing evidence. “No evidence” is for cases where none would be expected.

Of note is that item 11 essentially encapsulates the entire disability determination in one question: “The ability to complete a normal workday and workweek without interruptions from psychologically based symptoms and to perform at a consistent pace without an unreasonable number and length of rest periods.” This is the essential question for the whole disability determination.

Using the 20 ratings as a foundation, the Reviewing Medical Consultant drafts a narrative in a section titled “Summary Conclusions.” This narrative is the documentation that is used by the Disability Analyst. The ratings are not considered by the Disability Analyst.

Medical Evidence

The initial disability determination is a case record review, a paper review. It is the responsibility of the individual who is claiming to be disabled to provide the SSA with medical evidence to support the claim. Medical evidence consists of clinical signs, symptoms, and/or laboratory or psychological test findings. Clinical signs are medically demonstrable phenomena that reflect specific abnormalities of behavior, affect, thought, memory, orientation or contact with reality. A psychiatrist or psychologist generally assesses clinical signs. Symptoms are complaints presented by the individual. The findings may indicate an intermittent or persistent impairment depending on the nature of the disorder.

Medical evidence also includes information from other informed sources, such as family members and rehabilitation therapists, who have relevant knowledge of the claimant's functional capacity and limitations. This information is germane to the assessment of the B criteria on the PRTF and for the MRFCA review. There are no SSA-mandated forms for the provision of medical evidence. The collection of medical evidence is initiated by the local SSA district office and continued by the state-level DDS reviewing team to the point at which a disability determination of either an allowance or denial can be substantiated.

If the sources of medical evidence identified by the claimant do not provide sufficient evidence necessary to make a disability determination, a consultative examination can be provided. The SSA or DDS pays to have the claimant interviewed and a report sent to the DDS. The Consultative Examiner is generally someone not known to the claimant.

The Listings of Mental Impairment, and the forms used by Reviewing Medical Consultants—the PRTF and MRFCA—constitute the medical aspect of disability determination. For claims that do not result in a medical determination (i.e., a denial at step 2 using the PRTF because the disability is not severe or an allowance at step 3 because the disability is so severe that it precludes work on a medical review alone), the Disability Analyst continues the review with additional nonmedical factors. It is the medical aspect of the review of claims for disability benefits based on mental disorders that received a scientific evaluation.


In 1984, prior to the publication of the 1985 Listings of Mental Impairments, the SSA, under the direction of the then-Assistant Commissioner for Disability Patricia Owens, contracted with the APA to design an evaluation of the soon-to-be-released standards and guidelines for the evaluation of mental impairments. The evaluation would include the Listings, the operational definitions for their implementation, and the forms (PRTF and MRFCA) that would be used in practice. The designed evaluation was accepted, and in 1985 the SSA awarded a contract to the APA to conduct the two-year evaluation.7 The following description is a summary of the study.

Scope of the Study

The study was designed to ascertain the accuracy with which the medical standards and guidelines used by SSA's medical consultants operationalized the statutory definition of disability due to mental impairment and their consistency with contemporary psychiatric knowledge and practice. Within this broad objective, the study sought to identify and characterize cases that were difficult to evaluate with the medical standards and guidelines and to pinpoint the specific source of difficulty, and suggest solutions (Pincus et al., 1991).


The study consisted of three components. The first investigated the compatibility of the SSA's revised medical standards and guidelines with the statutory definition of disability. This component provided the bulk of the study's empirical data. Component I employed 72 psychiatrists who were a demographically and professionally heterogeneous sample of APA's membership recruited from five geographically diverse cities. “Professional heterogeneity” meant the orientation or “school” of psychiatry represented, such as expertise in psychopharmacology or psychoanalysis, or experience with inpatient or outpatient, acute or chronic clients. Each psychiatrist was assigned to one of two study conditions. One—the sequential evaluation condition—received training and applied the SSA disability determination process and forms (e.g., PRTF and MRFCA) used by SSA's Reviewing Medical Consultants in actual case-work. SSA staff participated in the training. The second study condition—the statutory definition condition—reviewed claims on the basis of psychiatrists' knowledge of the characteristics and limitations associated with the disorders experienced by the claimants. Training for this study condition consisted of in-depth discussions of the statutory definition of disability, claimants' impairments, functional limitations, and whether or not the claimants would be considered disabled according to the law. None of SSA's forms were used by this study condition.

Psychiatrists in both study conditions were instructed on the use of the Clinical Disability Severity Rating (CDSR), specially designed for the study. This rating was designed to record the psychiatrists' decisions as to the degree of disability for work experience by the claimants. This was necessary for two reasons. First, the statutory definition condition needed to document a decision as to disability status. Second, the sequential evaluation condition needed to record a disability finding at the fourth and fifth steps in sequential evaluation.8 The CDSR is a 10-point scale from −5 (unable to work) to +5 (able to work) with no zero point, permitting a dichotomous rating with a negative score signifying inability to work. The CDSR was accompanied by a confidence rating that allowed the psychiatrists to scale the confidence of their CDSR judgments. Psychiatrists in the sequential evaluation condition scored claimants on the CDSR only after completing sequential evaluation and its forms.

The 36 psychiatrists in each study condition reviewed each of their assigned claims first as independent reviewers and then in a three-member panel within their study condition. Each panel was as demographically and professionally heterogeneous as possible. Each study condition had 12 panels. The rationale for the panels was that they would help participants make judgments reflecting the view of currently practicing psychiatrists.

The sample of initial claims for disability benefits based on mental disorders was provided by the SSA for the study. They had been filed by adults, had passed a quality assurance review by SSA regional offices, and were sanitized for confidentiality. One panel in both study conditions reviewed each claim.

The second study component was a survey of panelists to identify problems with the review process and to solicit suggestions about how to improve the review of claims.

The third study component was an in-depth narrative review of a subset of claims that had been rated differently by the two study conditions to understand the reasons for the discrepancy.

Major Study Findings

The overall study finding was that the proportion of agreement between panels of psychiatrists using SSA's process and forms and those using clinical judgment and additional discussion of the statutory definition of disability for each claim was 0.77, an agreement that was higher than chance (kappa of 0.46). The proportion of agreement and kappas within study conditions and at the individual level were similar and led to the conclusion that differences could be attributed to inherent difficulty of the claim, complexity of the sequential evaluation process, or the medical standards and guidelines themselves. Analyses further explored these sources of disagreement.

An analysis examined the agreement between study conditions when difficult claims were eliminated and found that the proportion of agreement between study conditions was 0.96 (kappa of 0.78), suggesting that problems in evaluation can be attributed largely to the inherent difficulty of claims. If the standards and guidelines led to markedly different disability judgments than the statute, then a far greater degree of disagreement would have been found between the two study conditions. Nonetheless, the findings do not rule out the possibility that improvements could be made to the standards and guidelines that would result in less difficulty and disagreement about disability status in the determination process.

Analysis of the Listings/PRTF

The next series of analyses examined the agreement between study conditions for each of the seven categories of mental impairment of the Listings as reviewed using the PRTF. Three of the Listings categories—Organic Mental Disorders; Schizophrenia, Paranoid and Other Psychotic Disorders; and Anxiety Disorders—were found to work well. Affective Disorders showed a statistically greater chance of disagreement about disability decisions than other disorders. Personality Disorders and Mental Retardation and Autism9 had low agreement rates but were not statistically significantly. The sample size for these two disorders was small. The sample size for Somatoform Disorders was too small to interpret its high disagreement rate properly.

Within the Listings, analyses explored the agreement for use of the A, B, and C criteria. Moderately high rates of agreement on the selection of which A criteria to adjudicate a claim were found for Organic; Schizophrenic, Paranoid and Other Psychotic; Affective; Mental Retardation and Autism; and Anxiety Disorders. There was less concurrence in panelists' selection of Somatoform Disorders and Personality Disorder as Listing categories in which to adjudicate a claim. Ratings of agreement for the B criteria were reasonably high in the aggregate, but less reliable for the individual B criteria. Additional analysis found the first three B criteria (activities of daily living, social functioning and deficiencies in concentration, persistence, and pace) to be valid whereas the fourth B criterion (episodes of deterioration or decompensation in a work-like setting) showed some reliability and validity but was weak overall.10

The C criteria for schizophrenia, paranoid, and other psychotic disorders were found to have low reliability but were valid. There were insufficient data to analyze the C criteria for Anxiety-Related Disorders.11

In Component III of the study, panelists reported that the period of review for the C criteria was confusing. It required a rating period of two or more years. Additionally, they noted that a large amount of inference was needed in rating the C criteria. For example, there is often no direct evidence of how claimants would function outside a highly supportive living situation if they have lived there two years or more. The clinical inference required to make this judgment as if the claimant did not live there was difficult.


Only the 20-item worksheet of the MRFCA was amenable to statistical analysis. Interrater reliability of the items was conducted. There were sufficient data to conduct two analyses. The first was whether there was any limitation for the given item (rating categories: “not significantly limited,” “moderately limited,” or “markedly limited” versus rating “no evidence of limitation in this category” or “not ratable on available evidence”). The second was the agreement among raters about the degree of limitation when a limitation existed. Reliability ratings were—at best—moderate. For some items, the reliability was so low as to suggest dropping the item. For most items, the reliability of degree of limitation was greater than deciding whether or not a limitation existed. This finding suggests that raters need further guidance in determining whether sufficient evidence is present.

Medical Evidence

During the independent reviews, each panelist rated each claim on the overall quality of the medical evidence in the claim folder on a five-point scale from “completely adequate” to “very inadequate.” The average rating fell between “mostly adequate” and “somewhat inadequate.” This finding must be taken in the context of the selection of claims for the study. They were not routine claims; SSA had selected them from a subset of claims based on their having passed a quality assurance for development of the medical evidence.

Although the survey of panelists in Component II of the study generally supported all of the above findings, panelists spontaneously reported that one of the major difficulties they encountered in the review of claims was: the lack of sufficient medical and other evidence; and the variability in the quality of medical evidence when present. They also noted that the chronological organization of the medical evidence based on the date of its receipt created difficulties in the review of claim folders.

Date of Onset and Period of Review

In Component III, panelists raised the issue of the difficulty in formally establishing the date of onset of the impairment as opposed and in addition to the date of onset of the disability. There is no place to document the clinician's decisions about the period under review on the PRTF. Without this documentation, psychiatrists discovered that they were rating different periods, sometimes leading to differences in clinical opinion, which resulted in consistent judgments when the timeframe was aligned.

Identification of Difficult Claim Types

For the purpose of allocating sources of discrepancies in order to home in on problems with the standards and guidelines, difficulties in the review process itself and difficult claim types were also examined. Difficult claims were those for which panel members disagreed about disability status. Analysis of difficult claim types was conducted in two ways. The first was to use study variables (e.g., panelist ratings of their confidence in their decisions and their ratings of quality of the medical evidence). These variables were used for the study's purpose.

The second approach was designed to identify factors that SSA could use on routine cases to predetermine which cases were amenable to a facile review and which might require more concentrated effort—perhaps a tiered review process: one process for “easy” claims and another for “difficult” claims.

The second approach was to use clinical and demographic variables found in routine SSA disability claims folders that would predict difficulty in the adjudication. The “clinical” factors that were included in the analysis were the presence of a notable physical disorder; the time of onset of impairment; work history; and notable alcohol or drug abuse. The demographic variables were age and gender. Three factors predicted difficulty regardless of the category of Listings/PRTF under which the claim was reviewed or the step of sequential evaluation at which disability status was determined:


presence of a notable physical disorder amongst claims based on mental impairment;


onset of impairment less than 12 months prior to review; and


being female.

In the SSA system, these claims would likely take extra time to review and be those—if denied benefits—that might be reversed on appeal.

Selected APA Recommendations

Based on study findings, recommendations were made in reference to areas in the standards and guidelines that warranted potential revision and to additional research that would further enlighten SSA's disability determination. All recommendations were made on the premise that the basic construct of the SSA's medical standards and guidelines for the evaluation of claims based on mental impairment should be retained. Recommendations from the APA study, described below, reflect those that are consonant with the focus of the IOM committee and are detailed to indicate areas for a research agenda.

First, the medical standards and guidelines for claims based on mental impairment can be improved by refinements of specific aspects. In particular, the study identified the following eight areas:


Listing/PRTF category for Affective Disorders;


the A criteria for Affective and Personality Disorders;


the B4 criterion (episodes of deterioration or decompensation in work or work-like setting that cause the individual to withdraw from the situation or experience exacerbation of signs and symptoms);


scale points for all four B criteria;


timeframe and instructions for C criteria;


extending C criteria to other episodic disorders (e.g., Affective Disorders);


onset of impairment and period of review; and


MRFCA items.

Second, the medical evidence upon which disability decisions are based needs improvement. A nationally standard form to be used by treating psychiatrists, psychologists, consultative examiners, and other clinicians should be designed and tested for collection of medical evidence for all claims based on mental impairment. Third, identification and special evaluation of difficult claims should receive consideration within the SSA system. Fourth, a systematic series of studies and research development activities in collaboration with other federal agencies, academic institutions and professional organizations should be developed and conducted. Three areas of endeavor were identified:


SSA should conduct a review of the study claims to identify whether or not the same medical standards and guidelines lead to different decisions within all levels of SSA adjudication.


A longitudinal study of the claims reviewed in the study should be considered to understand the natural course of those claims in the SSA system and the course of the disability status of those claimants.


SSA should investigate the use of panels to adjudicate difficult claims.

The APA provided additional detail about these recommendations and described other recommendations outside the scope of the IOM committee's interest. It is possible that SSA has conducted some of this work.

Because the premise upon which all APA recommendations were made was that the basic construct of SSA's medical standards and guidelines should be maintained for claims based on mental impairment, any changes, modifications, or refinements would have to be based on input that is conceptually compatible and scientifically robust. The WHO's ICF is such a resource. The next section describes the ICF in general and then as it specifically pertains to the SSA disability determination.


The ICF is one of the “family” of WHO classifications designed for use in a range of health and health-related applications. The classifications cover a broad range of health information issues and provide an international, standard language that enables communication throughout the world among various disciplines and sciences. For example, the WHO's International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10, or herein referred to as ICD) (WHO, 1993) is one of the most familiar. The ICD is a classification used to translate diagnoses of diseases and other health problems into an alphanumeric code for recording morbidity and mortality. The ICD has become the international standard diagnostic classification for all general epidemiological and many health management purposes. ICD and ICF complement each other.

The ICF classification covers a person's functioning and disabilities associated with health conditions at the body, person, and society levels. Functioning in the ICF refers to all body functions, activities, and participation as an umbrella term. Similarly, disability describes the impairments of body function, limitations in activities, and restrictions in participation. For activities and participation, it can capture both capacity and performance. The overarching aim of ICF is to provide a unified and standard language to serve as a scientifically based frame of reference for the description of health and health-related states. It is designed for use with physical and mental health conditions alike. Additionally, it specifically states that social security programs are among its potential applications. It will also provide the basis for a systematic coding scheme for health information systems that permits comparison of data across countries, health care disciplines, service settings, and time—data that complement ICD.

At a global level, three features of ICF compel further consideration by SSA. First, ICF is predicated on a universal model. The ICF is intended for use with all people, not a predetermined set identified as “the disabled.” It is designed in such a way that different program criteria and thresholds can be applied to it: it does not contain a standard criterion of its own.

Second, ICF was developed with an internationally cultural perspective that has direct application to the heterogeneous population within the United States. The ICF was created with cross-cultural application as part of the process. It was not developed solely in the industrialized North American and Western European cultures. Thus, ICF has applicability to the diverse populations in the United States.

Third, one component of the WHO family of classifications is the development of specialty-based adaptations. For example, the ICD-10 has a version for use in primary care settings. With sufficient interest, a special adaptation of ICF for work disability could be developed. The factors associated with work could be compiled, leaving the relative weights of associated factors and the threshold for any dichotomous (able or not able to work) or interval (percent of work disability) decisions to SSA's program criteria to operationalize.

The ICF was unanimously endorsed at the May 2001 World Health Assembly in Geneva by the Ministers of Health of its 190 member countries. The ICF was published in all six official WHO languages and launched in October 2001. Because of its recent publication, attention to its predecessor is warranted.

International Classification of Impairments, Disabilities, and Handicaps (ICIDH-1980)

In the late 1970s there was a groundswell of interest in the impact of ill health on a person's functional status particularly as it related to provision of appropriate health care. Advances in medical knowledge were preventing mortality from acute medical conditions. This resulted in more people living longer, but living with chronic health conditions and ensuing disabilities. The classification scheme of the first edition of the WHO International Classification of Impairments, Disabilities, and Handicaps: A Manual of Classification Relating to the Consequences of Disease (ICIDH-1980) (WHO, 1980) was created to offer a framework to facilitate the provision of health-related information notably for chronic, progressive, or irreversible diseases. The ICIDH-1980 posited a model of the sequence underlying the illness-related phenomena as disease leading to impairment, leading to disability, and on to handicap. ICIDH-1980 defined these three consequences of the health condition. Impairment is defined as any loss or abnormality of psychological, physiological, or anatomical structure or function; disability as any restriction or lack (resulting from an impairment) of ability to perform an activity in the manner or within the range considered normal for a human being; and handicap as a disadvantage for a given individual, resulting from an impairment or a disability, that limits or prevents the fulfillment of a role that is normal (depending on age, sex, and social and cultural factors) for that individual.

This landmark work was not without controversy. When juxtaposed against other disability models, such as that proposed by Saad Nagi, inconsistencies and irregularities became evident. This too was the finding of an IOM committee (Pope and Tarlov, 1991).12 However, ICIDH-1980 was the only model with an associated taxonomy.

ICIDH-1980 was accepted in many countries throughout the world. It is widely accepted and used in physical rehabilitation in The Netherlands. In France, the ICIDH-1980 definition of “handicap” became the basis for the national law upon which its disability benefits are based. Some research instruments based on ICIDH-1980 were developed, and a nascent body of disability research exists. The ICIDH-1980 was not generally accepted within the United States.

After a decade of use in services and research, the need for a revision became apparent. In the early 1990s the WHO undertook the revision of ICIDH-1980.

ICF Revision Process

The ICF revision process has four unique features intended to address identified shortcomings of ICIDH-1980 and to make it useful and applicable to the disablement experience in a range of relevant situations. First, ICF was developed as an iterative process. Initially, WHO worked with three collaborating centers and one international task force, the International Task Force on Mental Health, and Addictive, Behavioural, Cognitive, and Developmental (MH/ABCD) Aspects of ICIDH,13 to draft an Alpha version (with the acronym of ICIDH-2 at that time).

With this version, the WHO began to build up the second unique feature of ICF: to include input in the development of ICF from diverse cultures, languages, and geographical areas. The Alpha version was circulated for review and comment, and new collaborating centers were incorporated into the process. This aspect of the revision process breaks from the tradition of developing a classification in English, based on North American and Western European expertise and then translating it into different languages. The ensuing Beta-1 version (also with the acronym ICIDH-2), which incorporated feedback from the Alpha draft, was even more widely distributed internationally.

The third unique feature was a series of formal field trials that were designed to collect empirical evidence for additional revision. Included in the field trials were such queries as the cross-cultural applicability of the concepts and model underlying ICF and the meaningfulness and sensitivity of the ICF domains and items in different cultures throughout the world. Although field trials are not unique to the development of international classification systems (e.g., the ICD), they are unique to the ICF process. Beta-1 field trials obtained data from all major populated geographical areas and continents: North and South America, Europe, Turkey, Russia, India, Japan, Africa, and Australia. Beta-2 field trials were even more extensive.

The dissemination of the field trials highlights the fourth unique feature of ICF. The ICF is predicated on the biopsychosocial model that combines the best of both the medical and the social models related to the WHO definition of health. Participants in the Beta-1 and Beta-2 field trials included not only professional mental and physical care providers, but also administrators, advocates, family members, and people with disabilities themselves. The SSA participated in the Beta-2 field trials. Unlike ICIDH-1980, ICF is intended for use by the broadest of audiences in the array of cultures, professions, and people affected by disabilities.

As noted above, the U.S. Secretary for Health and Human Services, along with other Ministers of Health, endorsed the ICF as an official member of the WHO family of classifications at the World Health Assembly in May 2001.

Conceptual Model

In response to feedback, the ICF has been written in neutral rather than negative terminology. Advocates and people with disabilities in particular pointed out that the negative perspective of ICIDH-1980 and its terminology often were perceived as a negative description of the person with the disability rather than as a descriptive term about the disabling situation itself. Thus, ICF has employed neutral terms to the extent possible.14 As with the ICIDH-2, the ICF envisions three components: body functions and structure, activities, and participation. Body functions are the physiological functions of body systems including psychological functions. Body structures are anatomical parts of the body such as organs, limbs, and their components. An activity is the execution of a task or action by an individual. Participation is involvement in a life situation. Decrements or disabilities in these three domains are respectively considered impairments, activity limitations, and participation restrictions. Impairments are problems in body function or structure as a significant deviation or loss. Activity limitations are difficulties that an individual may have in executing activities. Participation restrictions are problems an individual may experience in involvement in life situations.

Contextual factors are a significant addition to the conceptual formulation of ICF. Contextual factors include both personal and environmental factors. Environmental factors are the physical, social, and attitudinal backgrounds of the person. They are external to the individual and may have a positive or negative impact on the person's functioning. Inclusion of the list of environmental factors is an innovation of the ICF. Personal factors, such as gender, age, and level of education, also have an important impact on the person's functioning but are not listed in the ICF.

The conceptual model of ICF is an interactive one. ICIDH-1980 was linear. Not only does the health condition have an effect on the function and structure of the body, on the limitation of a person's performance of activities, and on participation in society, but these three components of functioning can have a multidirectional impact on each other and on the health condition as well (Figure 1). Prior to the Americans with Disabilities Act of 1990 (ADA) people with ill health or with a disability were often limited in their participation in society. For example, an individual with a mobility disability might not have been hired for a job for which she or he was otherwise qualified. In this example, the ADA can be understood as an environmental factor that facilitates the reduction of barriers to participation in work.



. Interaction between the components of ICF.

The ICF is also seen as a dynamic, not a static, model. The components, domains, and items may be impaired, limited, or restricted either temporarily or permanently. Impairments, limitations, and restriction may be progressive, regressive, or static, and intermittent or continuous. The temporal quality is not fixed by ICF. The deviation from the norm may be slight or severe and may fluctuate in degree over time. ICF does not specify a threshold for the degree of deviation from the norm for a determination of disability status. It can be applied to legal definitions or program criteria and used to put them into operation.

The SSA's statutory definition of disability, some sequential evaluation components, and even the basic premise of SSA's disability benefits can readily be mapped onto this interactive conceptual model. For example, in SSA's definition of disability there must be a medically determinable physical or mental impairment that causes the inability to work. SSA's requirement for the impairments is that they either result in death or last at least one year; this is consistent with the ICF model. Additionally, the threshold for disability status can be accommodated by the model as well: it is designed to be applied to Social Security disability programs. In sequential evaluation, SSA considers the existence of jobs in the national economy; this is an environmental factor. Personal factors include gender, age, and education. SSA considers age, educational level, and work history. Disability benefits are essentially the replacement of earned income (stemming from SSDI) that facilitates participation in various domains of social life. None of these aspects of SSA disability program are inconsistent with ICF. It is intended for application to programmatic requirements.


WHO uses an image of tree, branch, twig, stem, and leaf to describe the taxonomic structure. The entire classification of functioning and disability is the tree, components are branches, domains are twigs, items are stems, and subitems are leaves. Although the core concepts underlying the three components (i.e., body functions and structures, activities, and participation) are distinct, when applied to daily life the boundary between the person's capacity and performance, and the environmental impact upon the expression of activities and participation, are virtually impossible to depict without being artificial or without an evidence base. Thus, two of the ICF's components are contained in one list of the taxonomy: activities and participation. Conversely, as noted above, the components of body functions and structure are separately classified.

The ICF is organized in one, two, and full levels of detail. Additionally, it contains an index. For the uninitiated, using the one- or two-level classification serves to familiarize the organization of the material and assists in locating relevant items. The full classification contains the definitions, inclusions, and exclusions.15

The operational definition permits ready understanding of the item for laypeople, for clinicians, and for researchers alike. The definitions apply across cultures and avoid trendy jargon, particular schools of thought, or specific professions.

Coding and Rating

To use this neutrally termed classification and to identify impairments, activity limitations, or restrictions in participation, each component of the ICF is accompanied by instructions for rating. Like the ICD, the ICF version has an alphanumeric coding system. The component, domain, item, and subitem are identified to the left of a decimal. To the right of the decimal, the coding structure contains space for qualifiers such as difficulty in performing activities or in being involved in life situations.

Impairments in ICF body functions are rated by one qualifier as the extent of the impairment on a five-point scale from “no impairment” (0) to “complete” (4), with mild, moderate and severe being 1, 2, and 3, respectively. This is generally thought of as a clinical rating, with the threshold determined by clinical expertise. Body structures have three qualifiers. The first is identical to body functions. The second qualifier uses nominal coding to describe the nature of the impairment. For example, 1 is used to indicate that there is a total absence of the body structure as in an amputated limb, 4 indicates aberrant dimensions of the body structure, and 6 denotes a deviation in the position. The third qualifier for body structure identifies the location of the impairment. For instance, 0 specifies that the impairment is in more than one region, 1 that it is on the right, 4 that it is in the front, and 7 that it is distal. Using all three qualifiers provides a rich description of the extent, nature, and location of the impairment of body structure.

Activities and participation have two qualifiers: performance and capacity. The first qualifier is performance, which demarks how much difficulty the person experiences in executing the task or being involved in a life situation as conducted in her or his usual day-to-day life. This means how and where the person spends time doing things. The second qualifier is capacity, which describes the person's ability to execute a task or action. Capacity is assessed in a formal test setting that is standardized to evaluate the true ability without the influence of the environment. Thus, capacity is considered to be evaluated in a neutral, standardized, or uniform environment. Variations of these two qualifiers indicate whether or not the person is using assistive devices or personal assistance in the assessment. Coding for both performance and capacity qualifiers is identical to body functions and to the first qualifier of body structures: it identifies the degree of difficulty.

In the structure of ICF qualifiers, there is room for consideration of additional qualifiers. It is anticipated that some users of ICF may develop their own qualifiers to suit their programmatic or other needs. For example, the SSA may wish to consider application of these or other qualifiers—or the development of an explicit algorithm using qualifiers—in refinement of the B criteria or revision of MRFCA.

Disability for Work

At the beginning of the ICF revision process, the International MH/ ABCD Task Force had a unique role. It was the only entity with responsibility for input into all aspects of ICF. Initially, each of three collaborating centers had responsibility for the first draft of either the then-labeled impairments, disability, or handicap sections.16 When the Alpha draft of the then-titled ICIDH-2 was compiled, all revision participants had responsibility for all aspects of the entire draft.

In the development of its first revision efforts, the International MH/ ABCD Task Force reviewed SSA's disability determination and the PRTF to ensure the inclusion of appropriate domains and items pertaining to disability for work. Particular attention was paid to the inclusion of the first three B criteria. The items that constitute the first two B criteria are currently located in the ICF activity component (i.e., activities of daily living [B1] and social functioning [B2]) while a major aspect of the third B criterion (i.e., concentration) is located in the impairment section (i.e., sustained attention). The description of the assessment of severity in the mental disorder listings includes extensive lists of both activities that constitute activities of daily living and social functioning, most of which can be found as individual items in the ICF. Because the ICF provides definitions, other parts of the third B criterion—persistence and pace—can be operationalized in one or more of the items contained under the ICF heading “carrying out daily routine” (ICF code d230).

Other aspects of work—responding to current thought in the field of psychiatric rehabilitation—were introduced under the activities component. These items include seeking, maintaining, and terminating jobs. Interpersonal aspects of work are also among the items in interpersonal relationships, such as “relating to persons in authority” and “relating with subordinates,” that can be applied to supervisory relationships. Other codes readily apply to relationships with coworkers. The ICF also contains a chapter on general tasks and demands that has items concerning undertaking single and multiple tasks, working independently and in groups, and handling stress. They are worthy of SSA's attention.

Related Assessment and Research Instruments

Research on three instruments related to ICF developed for assessment of, and research on, functioning and disability is in different stages. However, these instruments are worth review and consideration. Each is being tested internationally for its psychometric properties.

ICF Checklist The ICF Checklist is an instrument that provides an overview of a person's functioning and disability in the ICF structure. The current version is designed to be completed by a health care worker using any one or a combination of sources of information: written records; primary respondent; other informants; or direct observation.17 It is essentially an abbreviated list of the items of major interest in health and disability care from the ICF.

The first section asks about demographic information, including occupation and medical diagnosis. This is followed by Part 1a, which queries on impairments of body functions, and Part 1b for body structure. Part 2 contains a short list of activity and participation domains and items, Part 3 contains environmental factors and is followed by a narrative section designed to give a thumbnail sketch of the contextual factors—both personal and environmental—that might have an impact on the person's functioning. The first appendix contains a brief health information questionnaire and is followed by a second appendix that provides guidance to an examiner when interviewing a respondent for completion of the ICF Checklist.

Use of the ICF Checklist is based on the qualifiers found in the ICF rather than a dichotomous rating inferred by the name.

The ICF Checklist offers SSA a basis for a standard form for provision of medical evidence for claims based on all physical or mental conditions. It supplies a potential format for identifying the Listing under which a claim will be adjudicated, the domains in which the claimant is restricted in functioning, and the type of additional medical evidence that would have to be developed. The ICF Checklist could be used from the point of initiation at the SSA DO. The compatibility of ICF for this purpose, in fact, is apparent in that the first two sections of the ICF Checklist contain information that can be applied to and/or used to modify or replace other SSA forms. Because sufficient additional information is collected to assist in various aspects of the disability determination, the ICF format would prevent applicants from “gaming” toward a favorable disability determination. Furthermore, a modification of the ICF to make it more specifically suited to the SSA disability determination could be developed that would include the salient items in the domains of activities that have work-related items. Such a product might provide the basis for providing concrete information upon which to rate the B criteria at the DDS level of review.


In a separate but related activity in the revision of ICIDH and the creation of the ICF Checklist, the WHO and three institutes of the National Institutes of Health (NIH)18 joined in a cooperative agreement to develop disability assessment instruments for use in clinical settings and in epidemiological surveys. The WHO Disability Assessment Schedule II was designed for use in assessing disability irrespective of health-related etiology. It is calibrated for use to assess disabilities associated with mental as well as physical conditions.

The WHO DAS II is conceptually based on the ICF and queries six domains:


understanding and communicating;


getting around;




getting along with people;


life activities (work and household activities); and


participation in society.

The WHO DAS II has undergone testing of its psychometric properties, notably reliability, validity, and sensitivity to change. It underwent testing in approximately 28 centers in more than 18 geographically and culturally diverse countries. Three versions are the product of this endeavor: the full 36-item instrument; a 12-item screener; and a 5-item short form.

Because the WHO DAS II is designed to be used either in conjunction with a diagnostic assessment instrument (such as the Schedules for Clinical Assessment in Neuropsychiatry or the Composite International Diagnostic Interview) or as a stand-alone instrument, the rating of the domains of disabilities is preceded by two sections. The first section collects demographic and other background information.

Twelve screening questions are located in the second section. Like WHO DAS II, these 12 questions are intended to be equally relevant for people with either mental or physical health conditions. Preliminary studies at WHO have shown that this brief questionnaire can predict nearly 90 percent of the variance found in longer versions of WHO DAS II. Ongoing analysis of the WHO DAS II is being conducted that explores its relationship to disability for work. It might behoove SSA to consider additional research on this—or other brief screening instrument—to be used at the initial stages of the disability determination process to screen out claims based on slight or not-severe limitations in function related to work and to identify those with extremely severe disabilities. Research and data analytic strategies could identify the threshold for slight or not-severe and for extreme or so-severe limitations that would provide an evidence base for this step in sequential evaluation. It might be worth considering applying this standard even prior to the review of A criteria.


There are many compelling issues deserving of research. The committee's second interim report (Wunderlich and Rice, 1998) is replete with them. The research questions are important, timely, and utilitarian. This review of the SSA's disability determination of claims based on mental impairment, the APA study, and the WHO ICIDH-2 suggests additional useful, scientific avenues of investigation. However, at present there is no overarching strategy for identifying and prioritizing research necessary to improve the disability determination for SSA.

In the mid-1980s, mental impairments were added to the Listings, many of which take into consideration functional consequences of an impairment. The committee understands and supports the need to revise and update the Listings to restore them closer to their original purpose. However, the committee is not aware of any attempt to evaluate the currency and consistency of listings, or at least those groups of conditions that account for a significant proportion of the disability rolls. SSA appears to have made the decision to replace the current Listings with an index without any attempt to first evaluate the Listings and use the findings to update them or to guide in developing a new index. SSA should specify the desired levels of specificity and sensitivity and evaluate the current listings against those standards to serve as a baseline for creating the new index (Wunderlich and Rice, 1998, p. 19).

The APA study suggests that the medical component of sequential evaluation for claims based on mental impairment works sufficiently well that only refinements to SSA forms, identification of a period of review, and improvement of the medical evidence (e.g., development of a standard form for basic medical evidence) are warranted. The ICF with its cultural sensitivity and applicability to diverse populations is suggestive of ways of improving both the medical evidence and the SSA forms. These seem relatively small issues in the big picture—when or if the big picture is clear.

What is the big picture for SSA? What are the aspects of the determination process that work well and which are those that do not? Using SSA's existing data, data from the APA study, and findings of the National Academy of Social Insurance, it is possible to identify the major questions raised in the existing five-step sequential process and the proposed four-step sequential process:

  • Imagine a bell shaped curve (Figure 2)—a normal distribution of disability for work—among SSA claimants. (It is not truly a normal distribution. SSA data can supply this information.)
  • Bisect the curve with a perpendicular line. To the right of the line are those claimants who cannot work, who are disabled for work according to the SSA statutory definition. To the left are those who are not disabled for work.
FIGURE 2. Hypothetical distribution of disability for work among SSA claimants.


Hypothetical distribution of disability for work among SSA claimants.

The first step of claims for both SSDI and SSI benefits based on either mental or physical disorders in sequential evaluation decides whether the claimant is already working. These people are not processed beyond the DO and do not appear above. There are no plans to change the first step.

The second step in the existing sequential evaluation eliminates those with slight limitations (claimants at the far left on the curve). They represent an extreme and are easy to decide on medical evidence. This was not a problematic decision in the APA study. This step is accomplished using the B criteria on the PRTF. No such determination will be made in the proposed SSA revision. This group with slight limitations would be carried along the whole process. The Committee to Review SSA's Disability Decision Process Research in its second interim report recommended that SSA should use a global functional assessment measure to screen out people who do not have severe disabilities (Wunderlich and Rice, 1998, p. 22).

The third step allows benefits to those who are so severely limited (those at the far right of the curve) according to the B criteria of the Listings as rated on the PRTF. This too is a relatively easy group to identify. It was not considered problematic in the APA study.

One proposed revision to SSA's process would use an Index of Disabling Impairments to make this decision. The committee' second interim report (Wunderlich and Rice, 1998, p. 19) noted that the Index is supposed to be simple enough so that laypersons will be able to understand what is required and to demonstrate a disabling impairment. The Index is being or about to be developed. Yet, there is already a scientifically based finding that the B criteria of mental impairments review work well. Would it not be worth investigating whether those same criteria would work as well for claims based on physical conditions?

Next, disability determination sorts the dichotomous decision of ability or inability to work for the remaining claims in the middle. Without knowing the scientific justification of collapsing steps 4 and 5 of the present sequential evaluation, it seems a reasonable step on face value.

The difficult claims to adjudicate are those close to the line in the middle. This is the group in which false positives and false negatives are most likely to occur. False positives are those who can truly work but are erroneously allowed disability benefits. These individuals will not appeal. Are they likely to stay on the rolls? False negatives are likely to appeal. This is the step in sequential evaluation that the APA study identified as problematic. It is also the step at which SSA has considered using an individualized functional assessment. For claims based on mental disorders, this new assessment would replace the MRFCA.

In the existing sequential evaluation, steps 2 (slight) and 3 (so severe) appear to be quite separate. In fact, they are two decisions made at the same point, using the same information. After claims have had their medical condition identified by the A criteria, the medical evidence is rated using the B criteria. The ratings sort those with slight limitations from those who are very severely limited and those in the middle. This one part of the process actually makes three decisions: denials for slight limitations; allowances for very severe limitations; and those in the middle who need more thought. If the decisions are correct, it is a very effective and efficient step—the rating of the B criteria.

What makes sense is to have the DO staff conduct step 1 (is the person involved in substantial gainful activity?) and document the health condition or combination of health conditions that are causing the purported disability for work, and record the onset of the health condition and the onset of the period of disability, thus identifying the period of review. The SSA's form for adults, SSA-3368-BK, contains the applicant's report of these dates and is collected by the DO. Medical evidence can substantiate the health condition and its onset. This would eliminate a time-consuming review with A criteria.

Once medical evidence has been added to the case folder, it might make sense to turn to a brief functional assessment such as the WHO DAS II 12-item screening questions to eliminate the slightly and very severely limited claims as is currently done for claims based on mental disorders. This would settle the claims at both extremes of the distribution. Research here could investigate the applicability of the three B criteria to claims based on physical health conditions and foster a more robust set of B criteria and scale points

For the remaining claims, SSA could build on the APA study and look for the factors that predict difficult-to-adjudicate claims: coexisting mental and physical conditions; onset of less than 12 months; and female claimants. SSA can identify other characteristics from among the claims whose decisions are reversed upon appeal. Additionally, scores or patterns of ratings on the WHO DAS II 12-item screening questions might distinguish routine from difficult-to-determine claims. These will be the claims likely to be closest to the border of able or not able to work. The APA suggested that these difficult claims might be best served by a review by a panel of Reviewing Medical Consultants.

Information SSA Already Has

The answers to the following questions reside either in the various datasets of SSA or in reports submitted to it by such informed sources as the National Academy of Social Insurance. These data can identify the strengths and weaknesses in the existing process.


SSA can identify the distribution of disability among its claimants. It should do this for three groups: all claims; claims based on physical conditions; and claims based on mental conditions. SSA may suggest reasons why this should be done for both disability programs and SSDI and SSI separately for the three groups of claimants. The reason for conducting separate analyses for physical and mental conditions is to understand the generalizability of the APA findings to claims based on physical disorder. SSA may also consider separate analyses of other problematic categories of physical disorders.


SSA can identify the magnitude of importance for slight and very severe disability decisions. How many of these decisions are made? What proportion of claims do they represent? How many denials based on slight limitation are appealed and reversed? How many very severely limited people leave the rolls and return to work? How long does a claim take to assess when it is denied at step 2? How long does a claim take to adjudicate that is allowed at step 3?


What are the characteristics of claims that take a long time to adjudicate?


What are the characteristics of claims that are denied at steps 4 and 5 that are appealed and have the disability decision reversed?


Can SSA identify the types of medical evidence that are most difficult and/or take the most time to obtain?

The above information can begin to establish priorities for additional research. The information tells us what is working well and what is problematic. If it is working, why change?19

Nonetheless, with improvements in the submission of medical evidence, the number and percent of claims that can receive a disability determination at steps 2 and 3 are likely to be increased, thus reducing the number that continue in the determination process. Information based on ICF items might be requested as part of a standard submission of medical evidence.

The next issue is to sort the claims remaining after the slightly limited and very severely limited have been handled. For these claims, additional functional assessment—but no consideration of age, education, and work history—is planned in the proposed process. This is where new functional assessment forms or instruments are required. Research would have to explore these alternatives and the medical evidence needed for the assessment.

The above statements are consistent with Recommendation 4-1 made by the IOM committee in its second interim report (Wunderlich and Rice, 1998, p. 21).20

Problems Already Identified by Research

The lack of sufficient medical evidence and the low quality of the medical evidence that is provided were identified as serious issues by participating psychiatrists in the APA study. This was their impression even though the claims had been through an SSA review for the quality of the evidence. In other words, they may have been better or more complete than average claims adjudicated in the DDSs. In addition to the APA's recommendation that a standard form for medical evidence be designed and used nationally, other medical evidence collection procedures might be reviewed. For example, although SSA is seeking to streamline its process, something may be gained by collecting medical evidence in two stages: the first stage for slight and very severe decisions based on a modification to the first three B criteria, the second for more in-depth assessment of the remaining claims, notably those that are expected to be inherently difficult, near the border of able or unable to work. What is needed? Which items in ICF can guide the development of medical evidence forms(s)? Have some states developed standard medical evidence forms, and if so, how are they working?

The APA study also identified problems in reviewing claims based on mental disorder when a physical condition was present. This issue might be addressed by a de novo consideration of the confluence of review of claims based on mental and physical disorders—the basic requirements of work for the two groups do not differ. Can a useful set of work-related factors be identified across disorder groups? For example, the items in ICF include general tasks and demands that are the basis for work-related activities (ICF d210 through d299); communications (ICF d310 through d349); interpersonal interactions and relationships (ICF d710 through d7109, d740 through d7409); and tasks needed to find, get, and keep a job (ICF d845 through d8459). A review of the ICF along with SSA's regulations and relevant Program Operations Manual System might reveal additional items. Any rating of these items would be consciously made in view of the claimant's health condition and other criteria outlined in the SSA definition of disability.

How might the first three B criteria for claims based on mental disorders and their rating be approved and made applicable to people claiming disability based on physical disorders? Again, the ICF offers a rich resource both for the items and for rating the items. The fourth B criterion might be revisited along with the C criteria to create a method to acknowledge that certain individuals appear not to be disabled for work but are only compensated in their existing environment or accommodated to their current level of stress. A change to a work environment would destroy their fragile state.

Of note is that there are no analogous B criteria for the Listings for physical conditions. In general, functional restrictions in the Listings for physical conditions are limitations in the function of a body part, such as motion of a joint. Only in the Residual Physical Functional Capacity Assessment (analogous to the MRFCA) are there any physical functioning ratings that apply to all of the physical categories. Not all of them are limitations in activities such as lifting and carrying; some are limitations of specific body parts or systems, such as visual acuity. Many of the physical limitations are redundant with functions of body parts of the Listings of physical impairments (e.g., reaching in all directions [Residual Physical Functional Capacity Assessment] and range of motion in joints [physical Listings]).


Prior to developing a program of research, identifying issues for research, and specifying research mechanisms and methodologies, a strategy for establishing priorities needs to be created. At present, compelling research questions are vying for attention. The IOM committee may wish to consider establishing a working group to draft an outline of work that addresses the following three issues preliminary to creating priorities for research:


Using SSA information, identify weak points in the existing adjudication process (e.g., slight limitation denials) and identify strong points that can be built on or generalized across claims based on physical and mental disorders:

  • Develop a plan for “fixing” the sequential process. For example, if the first two steps work well, keep them. Then go to the third step to look at problem areas (keeping in mind that for claims based on mental disorders, steps 2 and 3 are not readily distinguishable as steps only for the decision that is made).
  • Acknowledging that there will always be inherently difficult claims, develop a way of preselecting them and creating a process for handling them (e.g., panel of reviewers).

Explore the differences between the adjudication materials for claims based on mental disorders and those used on physical disorders to see if the strengths identified by the APA study might not apply to the adjudication of physically disordered claimants, and attend to the weaknesses as well:

  • If the A criteria are to be eliminated as has been suggested at times, identify some method of documenting the association of the health condition with the disabled state.
  • Explore the factors that need to be included in the medical evidence and how the medical evidence will interface with SSA assessment forms.

Review the ICF, ICF Checklist, WHO DAS II, and WHO DAS II screening questions to identify which components can: be used to create a standard form for the submission of medical evidence; can be used as a screen of slight and very severe claims; and can provide the basis for a fuller, more detailed version of the functional limitation criteria for both physical and mental conditions, and conduct a thorough investigation of the scale points and ratings as they calibrate to disability decisions (not severe, so severe, and disabled for work) for both physical and mental conditions.

The product of this working group has the potential to lead to a cogent research agenda that allows the inclusion of additional important research topics. It will also make evident the type of research that needs to be conducted, such as clinical trials or instrument development research, and the methods and mechanisms by which this research should be conducted. It also might suggest partners in this important national and federal endeavor.


There are four overarching areas in which recommendations can be grouped: (1) disability determination; (2) identifying priorities; (3) conceptual, taxonomic, and assessment resources; and (4) research agenda. All of the following recommendations are empirically based and made with the proviso that each be accompanied by research in an iterative process to provide a scientific base to identify and substantiate improvements. These recommendations are based on the disability determination of claims based on mental disorders and on research of the process, standards, and guidelines used in the adjudication of claims based on mental disorder.

Disability Determination


Existing research on disability claims-based mental disorders finds that no major change to sequential evaluation or to the standards and guidelines used in this process (notably the Listings of Mental Impairments and the PRTF) is warranted—only refinements.


Because basic work requirements are independent of health conditions, the adjudication of functional capacity for work should be the same for claims for disability benefits irrespective of the type of health condition (i.e., physical or mental disorder).


The current process of sequential evaluation should be maintained with the following modifications:

  • There should be four steps to sequential evaluation: the current steps 4 and 5 should be combined.
  • The first judgment in the medical review should be identification of the date of onset and period of review.
  • As long as the health condition causing the disability for work (i.e., the A criteria of the Listings of Mental Impairment) can be identified, substantiated, and linked with functional limitations, the focus of these criteria can be shifted to a more thorough and substantive evaluation of functional capacity for work.
  • Based on an evaluation of functional capacity for work, sequential evaluation should screen out claims that are not severe and those that are so severe. These two disability decisions (i.e., denials for not severe and allowances for so severe) are currently steps 2 and 3 in the sequential evaluation. These two medical judgments of disability status are focused on the two extremes of the distribution of disability claims.
  • For the remaining claims, SSA should apply clinical and demographic factors that identify difficult-to-adjudicate claims and select these claims for review using a panel process that combines additional functional assessment and nonmedical factors as the last step in sequential evaluation.
  • For claims not identified as difficult to adjudicate, the final step in sequential evaluation should be applied, which combines additional functional assessment and nonmedical factors.

Identifying Priorities


SSA should analyze its existing data to identify and prioritize areas for revision and refinement.

  • SSA should compare the magnitude of its caseload and the decisions made at each step in sequential evaluation for all claims for disability benefits in the SSDI and SSI programs, all claims for disability benefits based on physical conditions in the SSDI and SSI programs, and all claims for disability benefits based on mental conditions in the SSDI and SSI programs.
  • SSA should identify and prioritize the weak points and strengths in the existing sequential evaluation for claims based on physical conditions and for those based on mental conditions by examining the magnitude of decisions made at each step in the process and by the proportion of denied claims that are successfully appealed. Steps 2, 4, and 5 can be evaluated in this fashion.
  • SSA can identify claimants awarded disability benefits at step 3 who leave the rolls and return to work.

Using existing data, SSA can identify the characteristics of claims that are difficult to adjudicate.

  • Based on the APA study findings, SSA can explore clinical and demographic factors that predict difficulty in reaching a disability determination.
  • SSA can identify inherently difficult claims by specifying the characteristics of those that take a long time to reach a disability determination.
  • SSA can examine the types of claims that are likely to be reversed on appeal to identify inherently difficult claims.

Conceptual, Taxonomic, and Assessment Resources

  • The WHO's ICF, an international classification of disablement and functioning, provides a culturally sensitive, research-based, rigorous yet flexible conceptual foundation for revisions to SSA's disability determination for claims based on mental and physical conditions. One of its specified applications is for Social Security disability benefit programs.
  • The ICF conceptual model is consistent with the definition of disability in the Social Security Act and with the process that puts disability determination into operation.
  • Domains and items in ICF are readily utilized in modifications to existing SSA disability determination forms in sequential evaluation. They would enrich SSA's conceptual development and assessment of functional capacity to work.
  • The ICF has related disablement assessment and research instruments that can be modified for use as standard forms for medical evidence and in sequential evaluation:

    The ICF Checklist can be used as the basis for the development of a universal form for submission of medical evidence. It should be reviewed for possible inclusion of additional items related to work activities.


    The WHO DAS II contains six domains, five of which relate to work as currently conceptualized in the Listings/PRTF B criteria applied to claims based on mental conditions. This instrument could be supplemented with additional physically exertional items and used as the residual functional capacity assessment for both mental and physical conditions in the last step of sequential evaluation.


    Additional research should be conducted on the WHO DAS II as adapted above, and on its scoring and scale point definitions, to provide a scientific judgment about work capacity. (At present, the part of the MRFCA form that is used in the determination is a narrative summary. The 20-item ratings are considered a worksheet.)


    The WHO DAS II 12-item screening questions is composed of two items from each of the six domains. The scaling of these items can be tested against the current B criteria thresholds for identification of the two extremes not-severe and so-severe claims. They can then be used for claims based on physical and mental conditions to make disability determinations at steps 2 and 3 of sequential evaluation.

Research Agenda


A working group of researchers from SSA, the National Institute of Mental Health, other federal agencies, and knowledgeable researchers from the private sector should be organized to develop a research agenda to

  • review the ICF, ICF Checklist, WHO DAS II, and WHO DAS II 12-item screening questions for needed modifications for applicability in SSA sequential evaluation;
  • refine assessment instruments for the evaluation of disability for work based on physical and mental conditions for use in sequential evaluation;
  • calibrate scaling of assessment to disability for work based on physical and mental conditions;
  • create a standard medical evidence form;
  • review SSA databases to analyze existing data and identify priorities for research; and
  • review IOM recommendations for research.

All refinements to the existing system should be based on research findings.


The development of revisions should be an iterative process with research findings providing an empirical base.


Appropriate research methodologies should be identified for the diverse research issues.

In summary, the evaluation of claims for disability benefits is a complex and difficult task. This task is also a small component of the many extrinsic factors that have bearing on the shape of the disability program, such as long- and short-range economic factors, the changing characteristics of the general population and labor force, and the general priority and ideology held regarding people with disabilities. This paper has limited its focus to the medical review of claims for disability benefits based on mental conditions and has been informed by the APA's evaluation of SSA's standards and guidelines used in disability determination. With the recent revision of WHO's ICF and the development of related disability assessment and research instruments, a new and valuable resource has become available for use in modifications of the tools used in sequential evaluation.

Building on the research database provided by the APA study, the development of the ICF as a conceptual model, and on a classification system of disablement and functioning and its related instruments, it is possible to suggest recommendations for a research-based agenda to refine sequential evaluation and the standards and guidelines that implement disability determination for claims based on both mental and physical conditions. This is made possible because basic work requirements are consistent across disorder types. Each of the recommendations that have been made is intended to be a component of an explicit research plan developed by an interagency working group.

Finally, it is important to note that all recommendations are made within the context of the SSA's definition of disability. There are no indications that any change to the statutory definition should be considered.


  • Americans with Disabilities Act. 1990. Public Law 101–336.
  • American Psychiatric Association. 1980. Diagnostic and Statistical Manual of Mental Disorders (3rd Ed.) Washington, DC: American Psychiatric Press.
  • Pincus HA, Kennedy C, Simmens SJ, Goldman HH, Sirovatka P, Sharfstein S. 1991. Determining disability due to mental impairment: APA's evaluation of Social Security Administration guidelines. American Journal of Psychiatry 148(8):1037–1043. [PubMed: 1853953]
  • Pope AM, Tarlov AR. 1991. Disability in America: Toward a National Agenda for Prevention. Washington, DC: National Academy Press.
  • Social Security Act. as in effect January 2001.
  • Social Security Administration. 2001. Disability Evaluation under Social Security. Listing of Impairments Adult [Online]. Available: http://www​​/professionals​/bluebook/AdultListings.htm [accessed September 24, 2001].
  • Social Security Administration Regulations. 20 CFR Ch.III 404.1520 [Online]. Available: http://www​​/cfr20/404/404-0000.htm [accessed September 24, 2001].
  • Social Security Administration Regulations. 20 CFR Ch.III 404.1520a [Online]. Available: http://www​​/cfr20/404/404-0000.htm [accessed September 24, 2001].
  • World Health Organization (WHO). ICF Checklist [Online]. Available: http://www3​ [accessed September 24, 2001].
  • WHO. 1980. International Classification of Impairments, Disabilities, and Handicaps: A Manual of Classification Relating to the Consequences of Disease. World Health Organization, Geneva.
  • WHO. 1993. International Statistical Classification of Diseases and Related Health Problems: Tenth Revision. World Health Organization, Geneva.
  • WHO. 2001. International Classification of Functioning, Disability and Health. World Health Organization, Geneva.
  • WHO [Online]. Available: http://www​​.html [accessed September 24, 2001].
  • Wunderlich GS, editor; , Kalsbeek WD, editor. , eds. 1997. Disability Evaluation Study Design: First Interim Report . Washington, DC: National Academy Press.
  • Wunderlich GS, editor; , Rice DP, editor. , eds. 1998. The Social Security Administration's Disability Decision Process: A Framework for Research. Second Interim Report. Washington, DC: National Academy Press. [PubMed: 25101458]



Cille Kennedy is a Policy Analyst at the Office of Disability, Aging and Long-Term Care Policy, Office of the Assistant Secretary for Planning and Evaluation, Department of Health and Human Services in Washington, D.C.


A claim “meets” the Listings if the condition is of such severity that it precisely matches the findings. If the impairment does not match but is clinically equivalent to and exceeds the severity of one of the Listings (as is required in “meeting” a Listing), it is considered to “equal” the Listing that it most closely resembles.


The most recent edition of the Listings of Impairments was published in 2001. No substantive changes were made to the Listings of Mental Impairment that have impact on this report.


The 1985 revision contained eight Listings of Mental Impairment.


The Listing for Mental Retardation contains four sets of criteria. The Listing for Substance Addiction Disorders contains nine. They are outside the scope of this review.


Although the Listings for both Mental Retardation and Substance Addiction Disorders do have C criteria, they are of an entirely different conceptual nature.


The Listings of Mental Impairments have been modified slightly since the APA conducted the evaluation. However, the essential study findings are still relevant.


The MRFCA does not lead to a medical disability decision. The MRFCA narrative must be combined by the nonmedical Disability Analyst with age, education, and work history.


Mental Retardation and Autism were combined as one listing in 1985.


The phrasing of the B criteria has been modified since 1985.


In 1985, these were the only two listings to have C criteria.


Because of the thorough review by this other IOM committee, the presentation of ICIDH-1980 here is brief.


The SSA had been a formal member of the International MH/ABCD Task Force since its inception. Upon endorsement of the ICF by the World Health Assembly, all task forces were disbanded.


For example, “work” is a neutral term. ICF retains that neutrality by classifying it under the component of “activity” (also a neutral term) but uses a negative rating of difficulty to assess limitations in the performance of activities.


The ICF is published in two versions. The short version contains the one- and two-level classification with the second level having definitions.


Impairments were developed under the auspices of the French Collaborating Center; disabilities/activities under the Dutch; and handicap/participation under the North American Collaborating Center (NACC). NACC is housed in the National Center for Health Statistics, Centers for Disease Control and Prevention.


A self-evaluation version is under development.


The three NIH Institutes are the National Institute of Mental Health, the National Institute on Alcohol Abuse and Alcoholism, and the National Institute on Drug Abuse.


This may be the information to which the committee alluded in its second interim report (Wunderlich and Rice, 1998, p. 13) in discussing the nature and extent of the problem with the disability decision process. The above information may provide sufficient information to act as a needs assessment.


The committee recommended that early in its redesign effort, the SSA should specify how it will define, measure, and assess the criteria it will use to evaluate the current disability determination process, as well as any alternative processes being developed.

Copyright 2002 by the National Academy of Sciences. All rights reserved.
Bookshelf ID: NBK220746


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (7.9M)

Related information

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...