This review belongs to a new series of reports, Closing the Quality Gap: Revisiting the State of the Science (CQG Series). The original CQG Series1 was commissioned by the Agency for Healthcare Research and Quality (AHRQ). The intent of the series is to assemble a critical analysis of existing literature on quality improvement strategies and issues for topics identified by the 2003 Institute of Medicine report Priority Areas for National Action: Transforming Health Care Quality.2 AHRQ was also charged with continuous assessment of progress towards quality and updating the list of priority areas. Subsequently, AHRQ identified people with disabilities as a priority population.

People with disabilities can present special care challenges. Medical problems for a person with a disability can be exacerbated or complicated by the presence of other medical, psychological, economic, and social problems. Optimal care requires coordination among those involved in various sectors, the goal being to maximize the individual’s function and quality of life. Since function, quality of life, and community integration are interdependent, coordination may need to span the care spectrum to address various elements of life (e.g., medical care and schools or social agencies). Coordination of care, with attention to the intersection of medical and social services, is congruent with recent policy attention on integrated care and medical homes.

This review examines how health care outcomes have been assessed for people with disabilities. Disabilities are often broadly categorized into general types: physical, sensory, or intellectual. They can also be differentiated by etiology/timing of onset: developmental disabilities, disabilities acquired through trauma, disabilities as manifestations of disease processes, and disabilities iatrogenically acquired from treatments for disease conditions. With aging, a person’s existing disability profile may change (aging with disability), while otherwise healthy people may develop disability as a consequence of aging (aging into disability).

Regardless of disability type or etiology/onset, outcome measures are essential for evaluating quality care. Outcome measures may be even more useful in combination with process measures that address the extent and success of care coordination. For example, a process measure might evaluate the level of coordination between educational and medical services or between social and medical services. In general, little attention has been devoted to this intersection between outcome measures and process measures. Further, only modest efforts have been made to develop customized measures.3,4

Outcomes may be expressed as elements that directly reflect a person’s status, such as quality of life or social functioning. They may also be more intermediate measures, such as access to care—a common problem for many people with disability. This report emphasizes the former.

Here we discuss outcome measurement issues and explore conceptual frameworks for thinking about measuring outcomes for research and quality improvement efforts. As with all frameworks that deal with complex concepts, categories, paradigms, or classes we present are at best “ideal types” rather than simple designations with clean boundaries. We conclude with a summary of the project scope.

What Is To Be Measured? Levels of Analysis

Examining outcomes requires a broad understanding of what is appropriate to be measured. The range of choices depends in part on the goals for the research or evaluation. The goals should drive the focus, content, and structure of the optimal measure.

We can address outcomes of care for people with disabilities from several levels. Table 1 illustrates the relationship between the level of focus and related salient questions. We distinguish interventions directed at a disability from specific interventions directed at a given medical problem for a person with a disability, or comprehensive programs designed to integrate medical and social services for people with disabilities. Particularly for specific medical interventions not aimed directly at the disability, disability may often be thought of as a complicating condition that complicates the care (and changes the case mix), but for which the same outcomes apply as for people without the disability. However, some disabilities may impose a floor or ceiling on the outcome of interests. For example, problems with mobility that pre-date the disease may limit the potential for recovery or adaptation. A measure designed to tap deviations in performance for those expected to reach a near normal state may not be well suited for someone with an activity limitation.

Table 1. Levels of analysis for research related to people with disabilities, and related questions.

Table 1

Levels of analysis for research related to people with disabilities, and related questions.

What Is Measured (And Why)

Disability Definitions

It is challenging to define disability in a manner consistently applicable to the full human lifespan and range of abilities and activities. For the purposes of the National Healthcare Disparities Report, AHRQ used a definition that strives for consistency with the wide range of Federal programs related to disability: “People with disabilities are those with physical, sensory, and/or mental health conditions that can be associated with a decrease in functioning in such day-to-day activities as bathing, walking, doing everyday chores, and/or engaging in work or social activities.”3 At a recent AHRQ meeting, nationally recognized experts widely agreed that a single, consensus definition of disability is not feasible or desirable. Instead, they suggested that the definition should be governed by the research issue to which it will be applied.3

In the absence of consensus definitions, broad classifications can provide a useful alternative. Broad classifications offer a way to categorize outcomes for interventional or quality improvement initiatives for services for people with disabilities. Researchers have differed in their approaches to classification schemes. The medical approach focuses on pathology, such as classifications based on medical diagnosis, the body system affected, or functional loss or etiology of the disability. Psychological approaches tend to focus on the mental response to impairments. Social construction classifications distinguish between impairment and disability and tend to focus on discrimination against people with certain impairments, especially restricted access to services and opportunities.

Models of Disability

The different approaches to classifying disabilities reflect different models of disabilities. The medical model of disability5,6 emerges from Cartesian western medicine, which views treating or managing a disabling condition mostly as a matter of pathophysiology. That is, once the structural or biochemical deviations are understood, altering or controlling the disabling condition is then a compassionate and appropriate course of care. The biomedical model has been successfully used to guide the development of diagnosis and treatment. Due to this success, the model’s unintended consequences were not seriously examined until the latter half of the 20th century. One consequence of great importance was that defining health as the absence of disease equates the presence of disability with poor health. Another unintended consequence stemmed from the medical model’s narrow focus on solutions it could address, allowing for too little attention to nonmedical but important life outcomes including full participation in meaningful activity throughout the day and across environments (regardless of whether the underlying pathophysiology can be cured, controlled, or altered).

The social model of disability frames the problem as the societal response to the disabling condition, rather than the person. This approach separates the concepts of disability and health.5,7 The model holds that many of the disadvantages experienced by the person with the disabling condition are imposed by society, not inherent in the person’s physical or mental state. Responding to disability may then be as much or more a matter of political action than of health care per se. Hence, the disability community’s political agenda is based on a platform of civil rights, calling for equal access and opportunity.8

The social model acknowledges a person’s limitations but focuses on providing supports to enable the person to participate fully in family life, school, community, and work, whether or not certain skills or capacities are ever fully regained (or, in the case of a developmental disability, developed). This difference in focus is critical in the context of federal policies and programs focus on the medical or even the rehabilitative models. Such policies and programs sometimes deny funding for medical care to maintain a level of functioning. For example, people with certain developmental disabilities, notably cerebral palsy, face coverage limitations for occupational, physical, or speech therapy, because they are unable to achieve the standards of progress required for continued funding. Denied the therapy needed to maintain function, the person’s capacity to participate is thereby reduced. This problem particularly affects working age adults with disabilities who have aged out of school (and the medical and physical supports provided to children and youth in those settings under the IDEA legislation). There is some concern that physicians and other medical professionals who encounter children or adults with developmental disabilities focus on the disability and fail to provide standard care.9 For example, doctor or hospital visits for people with disabilities may not include standard preventative care if the medical professional focuses on the “problem” of the disability rather than on the person with the disability. Research reviews have consistently reported significant gaps in otherwise standard practices such as preventative dental care; mammograms; and routine screening for diabetes, high blood pressure, and other conditions (e.g., Larson & Anderson10).

The biopsychosocial model emphasizes how interactions between biological, psychological, social, and cultural factors affect one’s experience of health or illness.11,12 Some view this third model as an attempt to integrate the medical and social models.5 This view may represent the perspective of some allied health professionals, such as physical or occupational therapists engaged in rehabilitation. However, not all professional providers of social service-based support agree that the biopsychosocial model fully incorporates important aspects of the social model. For example, the social model differentiates between rehabilitation, which is designed to improve function or recover skill, and habilitation, which provides instruction to support a person’s skill acquisition throughout the lifespan.

The social and biopsychosocial models have influenced a number of Federal initiatives, such as the Americans with Disabilities Act of 1990, the New Freedom Initiative, The Surgeon General’s Call to Action to Improve the Health and Wellness of Persons with Disabilities of 2005, and Healthy People 2010. If people with disabilities are viewed as being on a health spectrum, similar to what people without disabilities experience, then well-being and health promotion are legitimate goals.13 In pursuit of these goals, Healthy People 2010 made important assertions: (1) that disability be treated as a demographic descriptor rather than as a health outcome (more discussion below); (2) that Disability-Adjusted Life-Years (which assume that disability signifies reduced health status) be eliminated; and (3) that disability be seen not as a static phenomenon but as a condition that varies by developmental phase, point in time, environmental context, and type of disability.14

Other efforts are ongoing to develop integrative and comprehensive conceptual models that will (1) acknowledge and work with the complexity of disability-related research and (2) provide the full healthcare spectrum. One example is a biopsycho-ecological model,15 which incorporates theories of Health Environmental Integration (HEI).16 The model adopts functional systems theory, viewing a person and her health condition as an outcome of a dynamic network of integral components including people, health conditions, and environment. “Health and illness occur within ecologies where small perturbations at any level (from molecular/cellular to environmental) can have large effects on overall person-level functioning and experienced quality of life” (Stineman, 2010, page 1036).15

Disability Paradigms

These models can inform and frame the perspectives of the range of professionals providing services for or engaging in research related to people with disabilities. Useful insights can emerge from considering how this informing and framing occurs. Broadly speaking, Figure 1 illustrates how the three general paradigms on care and support for people with disabilities overlap.

Figure 1 is a set of three overlapping circles illustrating the three general paradigms on care and support for people with disabilities. They are the social paradigm, the medical paradigm, and the rehabilitation paradigm.

Figure 1

Major paradigms of professionals that work with people with disabilities.

The social paradigm is common among professionals who (1) study people with disabilities and the effects of disabling conditions; (2) specialize in providing medical care to people with disabilities; or (3) focus on supportive services, including fields such as social work or special education. This paradigm acknowledges the appropriateness of medical and rehabilitative efforts specific to a particular person but emphasizes supporting and empowering people who have disabilities to be full participants in their families, communities, and schools, whether or not their disability or related medical conditions can be cured or fixed.

The medical paradigm is common among professionals who diagnose and treat people with disabilities and provide general medical care as well as specific treatment for the disabling condition. In this model, illness may be seen as a complication imposed on a person with a disability, or disability as a complication of treating a specific illness. Depending on the chosen specialty, a provider’s patient load may comprise only a minority of, or primarily, patients with disability. Curing is an ideal for which to strive. Both the medical and biopsychosocial models may inform the work of these providers. The relative weights for the models will be a function of personal concerns and professional training. Often the medical model strongly influences interventional research from within this paradigm.

The rehabilitation paradigm is common among professionals such as physiatrists and physical, occupational, or speech therapists. Patient populations include people with temporary disability due to trauma or illness, and people with “stable” disabling conditions. This paradigm strives to maximize function and optimize potential opportunities to participate in life as desired by the individual. Notably, the rehabilitation field is not cure focused, which is not the case for most other medical professions. Here as well, both the medical and biopsychosocial models may inform providers’ work. However, the biopsychosocial model, with its emphasis on person and environment factors, represents the dominant perspective that informs commonly used disablement frameworks.5

The different paradigms can play out in unintended communication problems as well. For example, the term “comorbidity” is variably used as a narrowly defined set of conditions that tend to travel together, such as diabetes and heart disease. Or it may be more broadly used as a synonym of multimorbidity, or any coexisting medical condition(s) that may complicate care for the medical condition of interest. However, co- and multimorbidity are medicalized terms and as such are often viewed as insufficient to describe the multidimensional and dynamic nature of a disability. To the extent that language shapes thinking, this impedes communication.

The figure presents the paradigms as overlapping because, as within any community, people use varying degrees of comprehensive or integrative perspectives. Some encourage or participate in bridging interdisciplinary work.

The Independent Living (IL) paradigm stands in contrast to the others. This paradigm proposes that people with disabilities are the best experts on their own needs and must take self-initiative in designing solutions for their needs. IL emphasizes the return of people to independent living and engagement in their larger community environments. People with disabilities are citizens first and consumers of health care second. IL is compatible with the biopsycho-ecological paradigm, both of which take into account the “full range of human experience in the world.”17

The Rehabilitation and Adaptation Perspectives

The course of addressing a disability can be divided into two basic segments: treatment/rehabilitation and adaptation. Each represents a distinct mind set about how a society responds to the challenges of disability. The relative size of these segments varies with the etiology of the disability. Three useful etiologies, or timings of onset, to consider are:

  • Congenital or developmental
  • Acquired (usually through trauma)
  • Aging
  • (Some would add a fourth category of iatrogenic)

The etiologies of disability affect the relative weight of the segments, in part based on life course issues. With developmental disabilities, for example, actions to address the underlying problem (e.g., surgery for spina bifida) are taken during a relatively brief period in early childhood (or infancy). Some people with developmental disabilities require medical treatment periodically as their bodies develop or their life circumstances change, or rehabilitation services at different life transitions (e.g., supported employment or job coaching). For some developmental disabilities no curative treatment is available. For most of the person’s life disability-related services are directed at helping the person with the disability develop skills and adapt his or her environment in order to foster the fullest participation possible.

People with acquired disabilities often face a prolonged period of treatment and rehabilitation. The goal, at least initially, is to restore the person to the same level of function as before the disability was acquired. However, for many, full restoration is not possible, and the emphasis shifts to coping with the remaining level of disability and preventing complications. Here again, emphasis is placed on adapting to a real living environment and actively participating in society. Thus, relevant outcomes include supported living, supported work, and full inclusion in the community.

For people with aging acquired disability, onset is linked to chronic disease and is often more insidious (although an acute event, like a stroke, may occur). There is an underlying belief that treatment is central; simply coping is often viewed as inadequate. This concept of disability is strongly linked to disease. As a result, it embodies an expectation that effective care will change the course of the disability and hence reduce secondary utilization of other services. In gerontology, a mainstay theory is that of selection, optimization, and compensation, which addresses a major coping strategy.18

The distinction among these etiologies of disability is not pure. People with developmental disability may be at greater risk of acquiring further disability because of limitations imposed by their primary condition. As larger numbers of such individuals survive into older adulthood, they are subject to age-related changes as well, and at younger chronological ages than the general population.19

People with disabilities, regardless of etiology, prioritize different components at different times. For example, parents of a child with a newly diagnosed disability often spend considerable time and energy seeking a cure or effective long-term treatment that will eliminate or greatly reduce the impact of the diagnosis on the child’s life course. In contrast, older children and adults with developmental disabilities prioritize getting support needed to live a fully included life, even if the underlying impairment cannot be cured or if function cannot be fully restored. For people with an acquired disability, an immediate effort to cure or full restoration of functioning is often followed by a prolonged period of rehabilitation and then a lifetime of getting support needed to live fully included lives. People with disabilities that result from degenerative conditions or the aging process often continue seeking curative or rehabilitative support until it is clear that death is imminent, at which time palliative care is often sought. Figure 2 illustrates the relative emphasis among these different types of disability.

Figure two is a series of six pie charts illustrating how the relative emphasis of medical care, rehabilitation, and adaptation changes for disabilities of different etiology and life stage.

Figure 2

Relative emphasis of medical care, rehabilitation, and adaptation for disabilities of different etiology. Note: Stylized examples to illustrate relative differences.

Here as well, thinking differs among professionals within different paradigms. Some social services professionals argue that this model illustrates one of the problems with services for older people. People receiving care for chronic treatments can continue to be active participants in their families, homes, and communities. Initiatives like the Center for Medicare and Medicaid and Services’ Money Follows the Person allow for more of the support people need to remain full participants.

These different conceptual etiologies have implications for the nature and extent of linkages with the medical care system. Care for people with developmental and acquired disabilities is directed at supportive services. Medical care is relevant only to the extent that the individual suffers from problems common to all people of the same age or from specific disease complications of the underlying condition. At the same time, disabilities may present access barriers to medical care (e.g., getting onto an exam table). Medical practitioners may need special knowledge about how to treat a given disease in the context of the disability. Successful care is generally measured using outcomes related to functioning and societal integration.

By contrast, older persons’ disabilities are more integrated into a disease framework. Treating the underlying disease is hard to extricate from treating the disability. Perhaps as a result of ageism, achieving societal integration is often viewed as less salient than improving the disease or disability level (or at least slowing decline).

As an illustration of these different mindsets, consider the following scenario. A disability activist who has paraplegia and uses a wheelchair is visiting his father, who just recently became a wheelchair user because of a stroke. In response to the nursing home staffs’ efforts to establish a program of timed toileting and ambulation training for his father, the son responds, “Forget that. Put an external catheter on him and let’s get on with life.” This also illustrates the occasional potential conflict between preferences (catheterize and “get on with life”) and clinical judgment and best practices (avoiding catheterization).

These distinctions have important implications for measuring disability-related outcomes. Table 2 illustrates some relevant outcomes. Those for developmental and acquired disability are virtually the same; both emphasize societal integration. In contrast, those for aging emphasize more limited functional goals, and indirect effects on costs and utilization of additional services.

Table 2. Examples of outcomes by major disability etiology.

Table 2

Examples of outcomes by major disability etiology.

Returning to the concept of disability as a complicating condition that hinders general medical care, the health goals of people with disabilities (as noted in the Outcomes column of Table 2) are not so different from those of the general population at comparable ages. The Institute of Medicine (IoM) framework proposed in their report “Crossing the Quality Chasm” (2001)20 outlines the basic goals for health care as follows:

  • Safe (no harm)
  • Effective (no needless failures)
  • Efficient (no waste)
  • Patient centered (no helplessness or unjustified routines)
  • Timely (no needless delays)
  • Equitable (no unjustified variation)

The IoM framework has been effectively used to distinguish healthcare for various subgroups, several of which are relevant to this discussion.21 Of the population segments outlined in the subgroups (given in Table 3) the groups most like the populations relevant to this review are subgroups 5 and 8. However, the population segments obscure the important distinction between disability and health. For example, a person living with a mobility disability may simultaneously inhabit more than one segment, perhaps having a “stable disability” but also a concern with maternal and infant health, or an acute illness unrelated to the disabling condition. A person with an intellectual disability may be healthy and have the same priority concerns and goals for health care as subgroup 1.

Table 3. Population segments and health priorities.

Table 3

Population segments and health priorities.

The “Aging Into” and “Aging With” Perspectives

Another recent major advance in understanding outcomes arises from applying a life course perspective. As noted, many people age into disabilities through the advent of illness. Many serious developmental or acquired disabilities have attenuated life spans. With improvements in care, many more people with significant disability now live to reach much older ages.19 Specific consequences vary by disabling condition, but in general this group may manifest age-related conditions at a younger age.2226,27 Thus, distinguishing the issues attributable to the underlying disability from those associated with aging presents a special challenge. In this way, underlying disability accentuates a long-standing geriatric dilemma; that is, to identify the etiology of a problem as attributable to pathologic change or a normal aging change. This blurring may have more important implications for diagnosing health concerns in a timely fashion.

The Individual’s Perspective

From the individual’s perspective, the concept of health is dynamic, sensitive to the conditions present at the time it is measured.13 Participants with disabilities use different criteria and themes when rating their health compared to those without disabilities. For example, four major health themes emerged from 19 focus groups with people with disabilities: (1) the ability to function and the opportunity to do what you want, (2) independence and self-determination, (3) an interrelated physical and emotional state of well-being, and (4) being unencumbered by pain.28 Compared to people without disabilities, these focus groups stressed the importance of resilience and emphasized their experience of health as distinct from their disabilities.

The paradigms and perspectives discussed above find traction in how the relevant outcome domains are examined and measured. In the next sections we discuss issues related to outcome domains and measurement. This is followed by a brief discussion of the World Health Organization’s international classification system for outcomes to assess health and functioning.

Relevant Outcome Domains

Outcomes relevant to people with disabilities encompass more domains than are relevant for the general population. Along with the basic repertoire of condition-specific and generic outcomes measures, additional measures and methods may be required for assessing outcomes for people with disabilities (or for specific subgroups organized by type of disability).

Different camps within the disability research community disagree about the extent to which the outcomes of medical care should be assessed similarly for persons with and without underlying disability, especially developmental and acquired disability. Some view the outcome domains as similar to those applied to general populations. Essentially, they see disability as a complicating condition, to be included in an appropriate case mix correction, and argue that it does not necessitate the use of different outcome measures. In contrast, others hold that in addition to the outcomes measured for the general population, specific outcome domains and measures should be tailored to the populations of interest. This group advocates for more individualized approaches that include additional outcomes related to managing disability and preventing secondary conditions.

The latter camp also argues that quality outcomes for disabling health conditions do not address considerations directly related to disability.3 A committee of experts convened by AHRQ noted that “[c]ommon health conditions that can be profoundly disabling include some, such as diabetes and heart failure, [which] have quality measures that generally are widely accepted and used. Most of these quality indicators reflect processes of care (e.g., measurement of Hb [hemoglobin] A1c levels, ophthalmologic examinations, prescriptions for certain medications). These quality indicators do not address considerations relating to disability.”3 For example, treatment goals for a person with quadriplegia with an indwelling catheter may differ from those for a person recovering from hip replacement surgery. However, treatment goals for people with disabilities for conditions such as diabetes and heart disease may be very similar or identical to those for people without disabilities. So while health status may contain the same components for everyone, individuals may assign the components different weights based on their situations.29

Outcome measures can be generic or specific to the particular issue. Generic measures are useful to policy questions or large conceptual problems that address multiple populations and/or interventions or services. Condition-specific outcomes may still be multidomain but are designed for a particular population or situation. Condition-specific measures are usually more sensitive to change because they are more closely linked to the problem at hand. However, they limit the ability to compare across populations.

Recovery choice pathway measurement technique, where people are asked to select an optimal pattern of functional recovery that reflects their preferences, moves further toward the individual-level perspective.30,31 While recovery pathway methods are typically applied to rehabilitation, it may also influence medical care recovery for people with disabilities. Differences found in values assigned by different groups (such as patients or clinicians, or even subgroups within those) suggest implications for measurement goals setting, and how measures are scaled when addressing quality of life.

Health-related quality of life (HRQoL) focuses on how health status affects quality of life (as opposed to well being, which addresses the positive aspects of a person’s life). Quality of life is often measured with health status measures, such as the MOS SF-3632 that describe, but do not value, health. Other measures attempt to value health as a way to combine mortality and morbidity, leading to constructs such as the Quality-adjusted Life Year (QALY), or the Disability-adjusted Life Year (DALY). The EuroQual’s EQ-5D is a widely used HRQoL measure. Summary utility scores have also been mapped for SF-36 data, allowing the SF-36 to be used to generate QALYs.33 These approaches place an immediate ceiling on the potential benefit people with disabilities can achieve because their baseline status downgrades the QALY score. Using such measures in making policy decisions thus has substantial implications for people with disabilities.

Health and functional status measures that give no credit for the adaptations made to achieve functional outcomes by people with disabilities will lower their scores on certain measures. For example, the SF-36E (for “enabled”) was developed to accommodate people with physical disabilities34 and thus substituted the word “go” for “walk” or “climb” in questions regarding personal mobility. The words “walk” or “climb” were at best confusing, and at worst, offensive to people who use wheelchairs. Questions were also reordered from shortest to longest distance so that people did not have to repeatedly acknowledge mobility difficulties.

Outcome Measures in Research for People With Disabilities

Outcome domains shared with general populations may require a modified methodological approach for people with disabilities. Quality research uses measurement instruments to determine improvements (or lack of) in outcomes of interest. The characteristics of measurement tools should be considered, along with how they are used to assess the outcomes of care for people with disabilities.35 Characteristics to consider include:

  • Psychometrics (bias, validity, reliability, responsiveness) of specific measurement instruments
  • Availability of comparative norms and standard values
  • Measurement timeframe and the potential for fluctuating levels
  • Disability cut-points (if the measure is part of a general spectrum)
  • Does the disability cut-point create a potential for floor or ceiling effect?
  • Types of patient-centered measures
  • Modes of administration and respondent burden
  • Data sources

People with disabilities have been disadvantaged in participating in research studies. Standard research instruments are not accessible to people with disabilities.34 For example, standard telephone sampling methods can miss those who cannot reach a phone by the 10th ring, or those whose primary mode of communication is not speech. Similarly, surveys that do not allow proxy response very often exclude people with intellectual or other cognitive disabilities. Most unfortunately, the concepts and wording of some health surveys are insensitive and even offensive to the firsthand experience of people with disabilities.34

Accommodation and universal design are two approaches to providing access to research participation. Accommodation requires enabling the measurement tools and modes of administration to allow access to people with disabilities. The SF-36E is one example of a tool adapted to provide accommodation.34 Universal design is being extended to the field of health services research.34 Universal design strives to develop products that are usable by all people, to the greatest extent possible, without adaptation or specialized design.36 The NIH’s PROMIS (Patient-Reported Outcome Measurement Information System) initiative is developing data collection tools, including computer adapted testing, based on the principles of universal design, i.e., they are applicable to persons with and without disability.37

Appropriate outcome measures may not differ between disabled and nondisabled populations, but the methodological approach to assessing outcomes may require accounting for patient characteristics or case mix. Of interest are the independent variables relevant to accurately reflect outcomes.

When not addressing disability as a direct cause of a medical condition, the relationship between disease and disability can be examined in terms of two different underlying paradigms. In one case, disability is a complicating condition acting as a confounder in elucidating the relationship between the treatment and the disease outcome. One technique to address this issue is treating the disability as a demographic descriptor, as suggested in Healthy People 2010. An example of this procedure can be found in a study analyzing access to care for people with various types of health care coverage in the National Health Interview Survey – Disability Supplement. In this analysis, outcomes such as short hospital stays and days of activity restriction in the previous 2 weeks varied with age, gender, race, overall health status, disability type, and health coverage status.38 Alternatively, the disability may affect either the choice of treatment or the effectiveness of that treatment. For example, the design of a physical activity regime for an adult with uncontrolled diabetes will likely be different for people with or without a significant mobility limitation. In that instance, the disability must be analyzed as an interaction variable.

Outcomes such as independent completion of activities of daily living (ADLs) have been shown to fluctuate widely over time. A person’s ADL level cannot be assumed to be stable.39,40 This variation imposes substantial problems in assessing treatment effects. At minimum, it implies a measurement error. It makes it more difficult to assess the extent of change. If the treatment actually changes the amount of variation in the ADL (e.g., stabilizes it), then its effect may be missed unless the research design is sensitive to the variation and magnitude of change.

ICF Outcome Domains

The International Classification of Functioning, Disability, and Health (ICF) was created as a framework to classify and assess function and disability associated with health conditions.41 The framework (Figure 3) rests on a positive description of human functioning rather than emphasizing the negative consequences of disease. This multidimensional model incorporates several levels of functioning. The framework attempts to explicitly acknowledge the dynamic nature of disablement, which can fluctuate based on a number of contributing factors across an individual’s life course.

Figure 3 is a flow diagram depicting the international classification of function, disability, and health framework. The framework identifies three levels of human functioning. The first level, health condition, designates functioning at the level of the body or body parts. The second level designates functioning at the level of the whole person. The third level designates functioning of the whole person in their complete environment. Within the whole person level are three domains of human functioning: body functions and structures, activities, and participation. The body functions and structures domain involves the physiological functions of the body systems, and the anatomical parts of the body.

Figure 3

The International Classification of Function, Disability, and Health (ICF).

The framework identifies three levels of human functioning. The first level, health condition, designates functioning at the level of the body or body parts. The second level designates functioning at the level of the whole person. The third level designates functioning of the whole person in his or her complete environment. The whole-person level includes three domains of human functioning: body functions and structures, activities, and participation. The body functions and structures domain involves the physiological functions of the body systems and the anatomical parts of the body. Impairments are problems with the body function or structure that result in a significant loss, defined as “deviations from generally accepted population standards.” The impairments may be temporary or permanent.

Activity and participation domains attempt to distinguish between the execution of a task or action by an individual versus participation in life situations. Later evaluations of the framework combined activity and participation into one category because of the difficulty in differentiating between the two domains.42 Contextual factors include environmental (external) factors that shape the lived experience of disability and can arise from the physical world, human-built structures, and from social and cultural constructions and attitudes. Individual experience is also shaped by personal factors (age, sex, coping styles, education, and other person-level differences in attitudes and behavior patterns). Since fewer personal factors are open to interventions, less effort has been made to elaborate coding for them.

The ICF has been criticized for lacking the clarity and distinction between the basic model components necessary for empirical measurement and testing.5 Subsequent empirical research on the concepts has suggested that the activity and participation domains were not distinguishable and has pointed to five potential distinct subdomains: daily activity, applied cognitive (higher order intellectual actions such as paying attention to multiple things at once, or following voice mail instructions), role participation, mobility, and social participation.42 Likewise, ADLs and IADLs map to self-care and domestic life chapters of the ICF, but other activities within those chapters did not map to ALDs/IADLs.

The last few years have seen some progress toward the clarity and discrimination needed for empirical measurement, and testing has been done in the last few years. While that level of analytical detail is beyond the scope of this review, some notable efforts are being made to establish useful ICF-based measurement.42 For example, the technique of functional staging combines the functional relevance of coding schemes (such as the ICF) with the ability to reliably measure an individual’s change, or discriminate between groups. Functional staging uses scale scores and has been tested for basic mobility.43

One commenter noted that the lack of a common conceptual scheme and language for disablement models has led to confusion within the scientific literature.44 The internationally developed ICF scheme congruently maps with disablement frameworks developed and employed in the United States.5 The ICF can help develop and promote a common language for improving collaborative, interdisciplinary, and international efforts. Further, the IoM has recommended its use in their report “The Future of Disability in America” (2007).45 Professionals involved in rehabilitation outcomes research suggest that the ICF framework will allow the sharing of outcomes across episodes of care that is necessary for evaluating quality coordinated care.46 A systematic review of ICF-related literature between 2001 and 2009 found evidence for the diffusion of ICF research and a growing trend in the use of the ICF framework. The authors concluded that the size and growth trend of the literature offered evidence of a cultural shift in the way disability and functioning are conceptualized and researched.47

Figure 4 provides a more linear illustration of the ICF to highlight how intervention points may differ for the “treatment” paradigms discussed above. Intermediate measures that assess the immediate effect of an intervention would likely vary based on the intervention point. Ultimately person-centered outcomes, such as quality of life, or living independently, remain somewhat universal to the intervention points. However, to be truly person-centered, the relevant outcome must relate to the goals as noted above.

Figure 4 provides a more linear illustration of the ICF to highlight how intervention points differ for the “treatment” paradigms discussed above. Intermediate measures that assess the immediate effect of an intervention would likely vary based on the intervention point. Ultimately person-centered outcomes, such as quality of life, or living independently, remain somewhat universal to the intervention points.

Figure 4

Adapted ICF framework. Note: Adapted from Colenbrander 2010

The ICF framework has limitations for assessing relevant outcomes. Since the ICF is focused on coding function at the person level, it omits system level outcomes that may be useful in evaluating quality of care or quality improvement initiatives. For example, cost and utilization numbers to examine use of second-order services noted in Table 1 above would not be gathered. Others suggest that tools developed by the World Health Organization that map to its ICF, the WHODAS II, could be useful identifying and aggregating information on utilization.39 However, the ICF is not encoded to include satisfaction or process measures used to assess the effectiveness of a new program.

The discussion on the characteristics of measurement above is, of course, equally relevant with regard to the ICF. There are multiple disability etiologies, classifications, definitions, and disability profiles. Thus, we must attend to the relevance and psychometric fit (potential floor or ceiling effects, relevant cut-points, and possibility of fluctuations in measures) of a given outcome measure to a given patient group.

The ICF provides a common disablement construct that is comprehensive in scope and complexity. However, we cannot predict the impact of the ICF on the assessment of outcomes of health services for which disability is most helpfully viewed as a complicating condition. How should prevention or treatment be provided and assessed for a medical condition for a person who happens to be experiencing a disability at that moment in time?

Key Questions

Our Key Questions (KQs) focus on the quality assessment component of quality improvement for specific interventions directed at a given medical problem not directly related to the disability itself, or comprehensive programs designed to integrate medical and social services for people with disabilities:


How are outcomes assessed for people with disabilities living in the community in terms of basic medical service needs?


What general population outcomes have been validated on and/or adjusted to accommodate disabled populations?


What types of modifiers or case-mix adjusters have been used with the general population outcomes to recognize the special circumstances of people with disabilities?


What are key parameters for measuring processes related to basic service care access for people with disabilities?


What measures have been used to assess effectiveness of care for people with disabilities living in the community in the context of coordination among health providers?


What measures have been used to assess effectiveness of care for people with disabilities living in the community in the context of coordination between community organizations and health providers?

The KQs were discussed with AHRQ, the CQG Series leader, participating EPC colleagues working on related projects, and members of the Technical Expert Panel (TEP). Discussions addressed the scope of this project in light of the structure and organization of the larger CQG Series and how this review could contribute to this area of research in disabilities.

Project Scope

The project scope included community-dwelling people of all ages who were diagnosed with and/or documented to have physical and cognitive disabilities. As a first cut, scope was confined to efforts contained within the Department of Health and Human Services; thus vocational rehabilitation was not in scope. Severe and persistent mental illness (SMI) as a primary diagnosis was determined to be outside the project scope. The disability profile of people with SMI, and in particular the cyclical nature of SMI, suggest that some of the processes and outcomes needed for this population would be qualitatively different. Service settings within scope included outpatient health, home, and community-based services.

To keep the project scope within feasible bounds, we focused on specific interventions for a medical problem for persons with disability or comprehensive programs designed to integrate medical and social services (from Table 1 above). The other levels of analysis (research on impact of public policy or geographic variation, the effect of organized programs, or interventions for the disabling condition) were outside the scope.

Within the level of individual outcome analyses, the quality improvement rubric informed the areas of interest. We chose a finite set of medical services and associated outcomes to represent conditions experienced by people with disabilities most likely to be considered prime targets for future quality improvement initiatives. We also focused on coordination of care as a major component for potential quality improvement initiatives. These scope limitations are discussed in greater detail below.

Illustrative Medical Conditions

We further narrowed the review scope to a meaningful but manageable set of medical services and associated outcomes. Our aim in so doing was to provide an illustrative set of healthcare encounters not specifically related to treating the disabling condition or eliminating related impairments. This finite set was made meaningful by choosing several basic medical service needs. We defined basic medical service needs for this project as preventive dental and medical care.

We also addressed a set of secondary conditions common to people with disabilities. A secondary condition is “any condition to which a person is more susceptible by virtue of having a primary disabling condition.”49 Preventing secondary conditions is identified as an important goal.45 Research identifying secondary conditions is growing,5054 including efforts to validate an instrument that identifies secondary conditions using ICD-9 codes.55 For the purposes of this report, we chose secondary conditions that also mapped to ambulatory care-sensitive conditions. Secondary conditions that are also ambulatory care-sensitive conditions represent prime targets for future quality improvement initiatives. The list of basic medical service needs and secondary conditions for this review includes:

  • Preventive dental care
  • Preventive medical care (based on general recommendations of the U.S. Preventive Services Task Force)
  • Urinary tract infections
  • Pressure ulcers
  • Diabetes and diabetic complications
  • Pneumonia
  • Asthma
  • Gastroenteritis
  • Hypertension
  • Obesity

Care Coordination

No consensus definition exists for care coordination, even as it is pursued as an important domain of quality care.56 A broad definition derived from a systematic review of care coordination within health care describes care coordination as “the deliberate organization of patient care activities between two or more participants (including the patient) involved in a patient’s care to facilitate the appropriate delivery of health care services. Organizing care involves the marshalling of personnel and other resources needed to carry out all required patient care activities and is often managed by the exchange of information among participants responsible for different aspects of care.”57 Coordination of care may extend beyond health care services to encompass other services for people with disabilities.

Care coordination is a multidimensional construct.56 What is successful care coordination depends on the perspective, whether from the patient/family, provider/professional, or system-level. A number of frameworks to describe care coordination and facilitate related research have been developed.4,56 One framework cites coordination activities as including:56

Establishing accountability or negotiating responsibility, including that of the patient and patient involvement

  • Communication
  • Facilitating transitions
  • Assessing needs and goals
  • Creating a proactive plan of care
  • Monitoring, following up, and responding to changes
  • Supporting self-management goals
  • Linking to community resources
  • Aligning resources with patient and populations needs

To these we add assuring that patient preferences are addressed and that the patient plays an active role in his/her care.

Broad approaches often used to carry out these mechanisms include: teamwork focused on coordination, health care homes, care management, medication management, and care coordination enabled by health care information technology. The mechanisms and broad approaches delineated above are expected to change as the general knowledge base expands.56 A taxonomy of quality improvement strategies was developed for the original CQG Series.1 These quality improvement strategies are:

  • Provider reminder systems
  • Facilitated relay of data to providers
  • Audit and feedback
  • Provider education
  • Patient education
  • Promotion of self-management
  • Patient reminder systems
  • Organizational change
  • Financial, regulatory, or legislative incentives

To these approaches we also add process measures that capture evidence of enhanced patient/provider communication, and patient-centered care.

Other than incentive structures, the remaining eight strategies are potentially relevant to the coordination of care for people with disabilities, including potential links with other human services agencies.


Our search focused on outcomes, patient experience, and care coordination process measures because they were immediately salient to exploring the interface of medical care and disability. Given the review focus on the individual level of analysis discussed above, relevant outcome measures are at the individual rather than population level. Examples of patient-centered outcome measures include functioning, psychosocial adaptation to disability, community participation, and social relationships. Measures related to performance of care coordination were also of interest.

In keeping with the perspective of disability as a complicating condition, we focused on generic outcome measures for the general population or for broad classes of disability. An alternative approach, searching for condition-specific measurement tools, was either too burdensome if all disabilities were included, or the applicability of the review was too restricted if only a few example disability conditions were included. Likewise, detailed mapping of the outcomes to the specific coding within ICF chapters was outside the scope of this review.