Health Disparities

The International Classification of Functioning, Disability, and Health (ICF) describes body functions and structures, activities and participation, environmental factors, and personal factors that interact to influence a person's function and disability.⁴ Within the ICF framework, health disparities research examines how differences in health activities, health care participation, and health outcomes relate to differences in body function or structure, personal characteristics (such as age, sex, race, sexual orientation, health conditions, fitness, life experience, individual psychological assets, education, socioeconomic status), or features of the immediate (settings such as home, workplace and school) or societal (such as social structures, services, social networks, laws, rules, attitudes, and ideologies) environment in which a person lives.

People with disabilities experience many health disparities. Some documented disparities include poorer self-rated health; higher rates of obesity, smoking, and inactivity; fewer cancer screenings (particularly mammography and Pap tests); fewer breast conserving surgeries when breast cancer is diagnosed; and higher rates of death from breast or lung cancer.⁵

Health disparities research has developed through four generations: (1) documenting the disparities, (2) exploring possible reasons for the disparities, (3) providing evidence for solutions, and (4) moving towards structural, multi-level interventions.⁶ This review focuses on studies that test interventions to reduce health disparities (third and fourth generation disparities research). However, disability health disparities research is largely first generation,⁷ focused on accurately documenting the health care disparities experienced by its diverse subpopulations. Documenting health care disparities is difficult for many reasons, including the presence of multiple disability subpopulations and ways of defining these subpopulations, and the lack of national surveillance data for specific subpopulations that results in many small, convenience sample studies.⁸

Cultural Competence

Cultural competence has been widely promoted as one approach to reduce health disparities. Since cultural competence remains variously defined and operationalized, it has become a blanket term to describe a broad range of system- or provider-level interventions. Initially, cultural competence focused mostly on racial and ethnic differences. More recently, it has been expanded to other marginalized population groups who are at risk for stigmatization for reasons other than race and ethnicity and/or who have differences in health care needs that result in health disparities. People with disabilities comprise some of these other populations. Culture competence implies the existence of a shared culture. Gill described the values of disability culture to include: viewing interdependence and helping as typical; accepting diversity in all its forms; being able to laugh at ones situation; planning for multiple potential outcomes and maintaining the ability to adapt; and being able to read people well under various conditions.⁹ Eddey and Robey described professional competencies related to the culture of disability, including: communicating with patients who have verbal deficits; understanding the values of people with disabilities and of disability culture including interdependence; and encouraging self-advocacy for patients and families.¹⁰ The authors acknowledge that, similar to other populations that experience health disparities, the degree to which a person identifies with disability culture varies and depends on a number of factors, including: the stage of life in which a disability is acquired; the visibility and/or severity of the disability; and the amount of organizational support for disability culture in the area where one lives.¹¹ The interventions described in this review attempt to reduce disparities for persons with disabilities by affecting change in providers or health care systems. This review tries not to take a position on the concept of disability culture.

Scope and Key Questions

PICOTS

Table 2 provides the populations, interventions, comparators, outcomes, timing, and settings (PICOTS) of interest. The analytic framework can be found in Chapter 1 and Appendix A.

Table 2

Review PICOTS—disability populations.

Literature Search Strategy

We searched Ovid MEDLINE®, PsychInfo, and Cochrane EPOC from 1990 to June 2015. Because the concept of cultural competence interventions in the disability community is not well defined, the initial search cast a wide net into the disability literature. Searches were performed iteratively to identify concept boundaries and tighten the working definitions and eligibility criteria to balance search sensitivity and specificity with feasible numbers of references to screen. The initial search strategy included cultural competence specific intervention terms, adapted from the race/ethnicity literature, and more general terms targeting health accessibility and health promotion. These more general terms were included because the exemplar articles identified by members of the technical expert panel were focused on improving physical access to care for individuals with disabilities and reminding providers to see the “whole person” with regard to providing preventive care and care for comorbid conditions. The health promotion, primary prevention, and health service accessibility terms had the greatest yield and were, therefore, the focus of the revised search. Other terms included in the revised search highlighted attitudes of health personnel, health care disparities, and the presence of stigma. The final search algorithms are provided in Appendix B. We also manually searched reference lists from systematic reviews and used back and forward searching of key articles recommended by experts.

Study Selection

We reviewed bibliographic database search results for randomized controlled trials (RCTs) systematic reviews, nonrandomized controlled trials, before and after observational studies with comparators, and interrupted time series studies published in English language relevant to our PICOTS framework. All studies identified by title and abstract as potentially relevant by either of two independent investigators underwent full-text screening. Two investigators independently performed full-text screening to determine if inclusion criteria were met. (Excluded articles at full text provided in Appendix C with reasons for exclusion.)

The full team vetted initial search results and adopted inclusion decision rules to clarify search results to address the review scope. This led to several refinements to the inclusion criteria. Patients with nonsevere mental health conditions, such as mild to moderate depression, did not meet our disability criteria; therefore, interventions integrating mental health services into primary care did not meet inclusion criteria. However, interventions targeting the attitudes of physicians toward people with mental illness, to the extent the condition qualifies as a disability, were eligible and included.

We had difficulty drawing tight boundaries around the interventions because the disability literature is often not identified by the terms “cultural competence” or “culturally appropriate.” Under our normative definition of cultural competence for this population, remote (e.g., web- or phone-based) medicine as an intervention was deemed culturally competent to the extent that it increases access for people whose disabilities make travel difficult. Although these interventions are not “clinic-based,” virtual interventions involving the formal health system (essentially replacing the need to travel to the doctor's office) create access in a unique way for the target population. These interventions are seen as conceptually parallel to infrastructure changes that improve access for people with physical disabilities.

Interventions aimed at improving physician or patient knowledge of existing treatment guidelines for conditions experienced by people with a disability are not included unless they also targeted physician perceptions and/or patient access to care. School-based interventions targeting the attitudes of teachers, classmates, and other professionals were excluded as outside the scope of this review, as were studies aimed at changing the attitudes of providers of long-term disability supports and services in community settings. Only studies that examined interactions with formal health care providers were included. As a result, several trials aimed at improving wellness and secondary disease prevention among people with disabilities in home and community settings, including some virtual interventions, were excluded.

Due to the limited available literature, we also expanded the criteria to include studies from other developed countries that tested interventions that could possibly transfer to U.S. health care.

In order to focus on the literature most likely to be informative, we also found it necessary to create decision rules for study comparators. Studies that used comparators that did not allow for direct testing of the cultural competence intervention/intervention component were excluded.

Differences of opinion regarding eligibility were resolved through consensus adjudication.

Risk of Bias, Data Extraction, Synthesis, and Presentation

We evaluated the risk of bias in included studies according to study design using criteria from the Cochrane risk-of-bias tool in interventional studies (Appendix D). Given the heterogeneity of the study populations and interventions and the high risk of bias assessment for most of the included studies, we were unable to draw meaningful conclusions from the literature. Because the literature did not provide adequate studies for quantitative analysis, results are summarized into evidence tables and qualitatively synthesized by common characteristics of interventions and outcome measures. Barriers to forming inferences from study results are also presented. One investigator abstracted the relevant data from eligible trials directly into evidence tables. A second investigator reviewed evidence tables and verified them for accuracy.

Literature Search Results

We identified 17,684 unique English language citations (Figure 4) from 1990 to June 2015. After excluding articles based on title and abstract, full texts of 80 articles were reviewed to determine final inclusion.

Figure 4

Literature flow diagram—disability populations.

Examples of studies excluded at full text for methodological reasons included: studies with a pre-post design or analysis,^13-15 studies with a weaker comparator population than included studies,^16,17 or studies where the intervention or outcomes were difficult abstract or interpret.^18-20

Twenty of the 25 included studies were RCTs; five were controlled trials.^21-25 Included studies fall into three main categories: interventions, predominantly trainings and curricula, aimed at changing professionals' attitudes towards individuals with disabilities (n=16); interventions aimed at increasing quality at the point of care by prompting patient and physician interaction (n=5); and interventions aimed at reducing barriers to accessing care (n=4). Table 3 describes the included studies by intervention type, disability population, and provider population. Studies were generally high risk of bias (Appendix D). Since the risk of bias and heterogeneity of the studies precluded any strength of evidence other than insufficient, we describe the studies by emergent patterns.

Table 3

Cultural competence intervention type by disability and provider populations.

Among the 16 studies aimed at changing professionals' attitudes through training or curricula, the majority focused on changing the attitudes and beliefs of medical (n=7),^{21,23,27,29,30,35,45} nursing (n=2),^22,26 pharmacy (n=3),^25,31,32 or physical therapy (n=1)³⁶ students. Three studies targeted practicing professionals: one studied primary care practice nurses;²⁴ and two studied mental health professionals.^33,34 Eleven of the 16 studies focused on physician attitudes toward people with a mental illness,^{21,25-27,29-34,45} three focused on attitudes toward people with a physical disability,^22,23,35 two studies focused on attitudes toward people with an intellectual disability,^23,24 and one study focused on attitudes towards people with lower back pain.³⁶ One study focused on a dually diagnosed population that had mental health and substance use concerns.³³

Five included studies attempt to reduce disparities at the point of care. Four of the five studies provide additional information about the patient and prompt physicians to pay attention to areas of known disparities during the routine clinical exam. One study teaches and empowers women with mobility impairments to get Pap tests and mammograms. All five studies promoting interaction between patients and physicians (or patients and the health care system) were conducted in the primary care setting; three focused on people with intellectual or learning disabilities, one focused on children with attention-deficit hyperactivity disorder, and one focused on women with mobility impairments.

The studies aimed at reducing barriers to accessing care were delivered by psychologists, occupational therapists, and masters prepared counselors. One of the four studies targeted people with mental illness, one focused on people with arthritis, and two focused on people with multiple sclerosis.

Interventions Aimed at Changing Health Professionals' Attitudes

The framing of these articles generally spoke to reducing stigma,^{26,27,29-31,34,45} changing stereotypic views,^21,33 filling gaps in training and understanding,^23,24 and familiarizing physicians with supportive equipment.³⁵ Table 4 summarizes the 16 studies of trainings or curricula aimed at changing health professionals' attitudes toward persons with disabilities.

Table 4

Summary of interventions targeting provider attitudes by disability type.

Table 5 summarizes the common characteristics and modalities of included interventions aimed at changing provider attitudes toward persons with disability. The majority of interventions included direct contact with a person with a disability. Of the 16 included articles in this category, 11 interventions included direct contact with people with disabilities,^{21-23,25,26,29-32,34,45} three included role playing with people with disabilities or actors playing the part of a person with a disability,^21,23,29 and two included a person with a disability as a trainer.^24,25 Twelve studies used one-time trainings or experiences of varying intensity,^{21,22,24-26,29,31-35,45} two developed university curricula,^23,30 and two tested different approaches to teaching subject matter (biomedical versus psychosocial models of lower back pain; and biogenetic versus multifactorial etiological framings of anorexia nervosa).^27,36 One older study used a simulation in which providers “assumed various disabilities” for a certain amount of time²² and one focused on medical students learning to use wheelchairs.³⁵ These last two simulation studies may be considered controversial or outdated methods of competency training by members of the disability community. Simulation exercises do not accurately portray the experience of having a disability and may have the effect of reinforcing negative stereotypes toward members of the population, and make disability an individual instead of societal problem.⁴⁶

Table 5

Aggregate of interventions targeting provider attitudes by disability type.

The form of the comparators varied. For the five studies that used no-intervention or wait-list control, the duration of the five interventions ranged from 4 hours to 4 years.^{22,23,29,33,35} Four studies that used the equivalent of a “usual care” control group generally embedded education or training components within curriculum that was otherwise relevant compared with the curriculum without the added component.^30-32,36,45 For example, one study compared a 12-hour Mental Health First Aid (MHFA) training program plus the standard mental health curriculum to the standard mental health curriculum alone for pharmacy students.³¹ Another study provided a 60-90 minute, contact-based experience as part of a mental health pharmacy course; students who had the experience later in the semester served as the comparator popuation.³² Two studies compared the effect of differing educational approaches, or disability framings, on student attitudes and beliefs.^27,36 Michaels et al. used a video attention control on an unrelated topic.³⁴ In addition to testing whether training is more effective than no training or usual care, four studies considered the comparative effectiveness of delivery modality^25,26 and intensity.^21,24 For example, one study had three arms: watching a DVD of individuals with mental health disabilities and their caregivers describe their experiences of mental health stigma followed by a researcher-facilitated discussion; the same testimonies delivered live, followed by a research-facilitated discussion delivered live; or a lecture on stigma provided by a mental health nurse researcher (no direct or indirect contact with consumers or care providers).²⁶ All three arms were roughly 75 minutes long.²⁶

Table 6 organizes the studies by outcomes studied. No studies examined clinical outcomes with a controlled design, although one of the included studies assessed nurses' self-reported changes in clinical behavior following the intervention to improve the accessibility of care for people with intellectual disabilities.²⁴

Table 6

Reported key study outcomes for trials aimed at changing provider attitudes.

Common outcomes included measures of stigma or attitudes,^{21-27,29-31,33,34,45} knowledge,^{21,24,29,31,33} self-efficacy or treatment confidence,^24,29,31 and intended social proximity or behavior (e.g., comfort working with someone with a mental health problem).^25,26,29,31 Scales used to measure stigma or attitudes included: the Attitudes toward Persons with Disabilities Scale,^22,35 Community Attitudes toward the Mentally Ill (CAMI) scale,²⁹ Opening Minds Scale for Health Care Providers (OMS-HC),^30,32 Mental Illness: Clinicians Attitudes Scale (MICA) scale,^21,26 Self-Determination Scale (SDS),³⁴ Recovery Scale (RS),³⁴ Causal Attributions Scale,²⁷ Attribution Questionnaire (AQ-9),^25,34 Opinions Scale,²⁷ Eating Disorder Stigma Scale,²⁷ and the Emotional Reactions to Mental Illness Scale (ERMIS).²⁶ Cutler et al. developed a 24-item instrument to rate the attitudes and feelings of medical students toward people with chronic mental illness.⁴⁵ Michaels et al. developed the Awareness Questionnaire (AwQ) to measure recognition of stigma in the mental health system and its impact on people with mental illness, and a stigma measure disguised as a knowledge test to circumvent social desirability.³⁴ Intended social proximity was measured by Reported and Intended Behaviour Scale (RIBS)^26,29 and the Social Distance Scale.²⁵ Knowledge was measured using the Mental Health Knowledge Schedule (MAKS).²⁹ The Jefferson Scale of Physician Empathy was used by one study.²⁹ Munro developed a knowledge questionnaire called the Comorbidity Problems Perceptions Questionnaire (CMPPQ).³³

Reported results were mixed. Twelve of the 16 studies measured physician stigma, attitudes, or beliefs.^{21-23,25-27,29,30,32-34,45} In six of the 12 studies, the intervention significantly improved attitudes/reduced stigma;^25-27,32-34 four of the 12 studies found no significant effect of the intervention on attitudes/stigma;^21,22,29 two of the 12 studies found both positive and negative effects of the intervention on attitudes toward people with a disabiltiy;^23,45 one study found arms to be differentially effective on provider attidues and stigma.³⁰ Knowledge was measured in six of the 16 studies.^{21,24,26,29,31,33} Interventions improved provider knowledge in three studies;^21,24,31 no significant effects on knowledge were found in three studies.^26,29,33 Treatment confidence was measured in three of the 12 studies.^24,29,31 Significant increases in confidence were found in two studies;^24,31 no significant effects on treatment confidence was found in one study.²⁹ Provider intended social proximity was measured in four studies;^25,26,29,31 half of the interventions showed significnat improvements on these outcomes.^26,31 The medium to high risk of bias and different PICOT combinations in the included studies does not allow for further analysis and synthesis of these reported effects.

Two trials reported outcomes that did not fit into Table 6. One study found students with the biopsychosocial lower back pain education had more positive attitudes and less fear in recommending general physical activity for people with lower back pain leading to more guideline consistent recommendations for work and activity.³⁶ In the other study, knowledge and skills measured were specific to wheelchairs.³⁵ Students in the intervention had statistically higher scores on the written wheelchair knowledge test and practical wheelchair skills examination compared with controls.³⁵ No difference in attitudes was observed between the intervention and control groups.³⁵

Interventions Prompting Interaction Between Patients and Physicians or the Health Care System

Table 7 summarizes five trials that prompt interactions between patients with disabilities and their physicians and health systems. Four of the five trials tested the effect of providing information or prompting clinician behavior at the point of patient interaction on the care received during that interaction.^38-41 One trial provided screening education tailored to women with mobility impairments and followed up the women to determine if care was received.³⁷ These interventions aimed to reduce disparities in provision of health or preventative care.

Table 7

Interventions targeting patient/provider interactions for disability.

Four of the five studies focused on individuals with intellectual or neurobehavioral learning disabilities.^38-41 The United Kingdom and Australia have implemented health checks for individuals with intellectual disabilities on a large scale. The strongest trial in this grouping is from Australia and tests the effect of a Comprehensive Health Assessment Program (CHAP).³⁸ The CHAP is a 21-page booklet divided into three sections: a medical history completed by the carer and carried to the doctor's office on the day of the appointment; a section completed by the doctor that documents he/she has reviewed the carer's completed medical history, completed a targeted examination, and developed a heath action plan with the carer; the last section has a list of commonly unrecognized or poorly managed conditions and disability-specific comorbidities.³⁸ Four hundred fifty-three people with intellectual disabilities participated in the cluster randomized trial. People with the intervention were significantly more likely to: have sensory deficits identified, receive a tetanus/diphtheria booster or hepatitis B serology, have their weight recorded, and women were significantly more likely to have a Pap smear and mammogram.³⁸ Many of the effect sizes were large. The United Kingdom and Australia have implemented health checks for individuals with intellectual disabilities on a large scale.

The other three studies that attempted to prompt physician behavior at the point of interaction had major theoretical and methodological limitations.^39-41 Unlike the CHAP tool that clearly prompted physician behavior (similar to checklist interventions), the Meurs et al. questionnaire provided a large amount of information without a pathway for action. The authors identified not giving dentists enough time to “digest” this information as a potential study limitation.⁴⁰ The intervention drew additional attention to the limitations of the person with the disability without assisting the dentist to identify strategies to change his or her approach to be more culturally competent. Further, this study offered no opportunity for patients and/or caregivers to rate the physicians on the interaction. The Turk and Wolraich studies had low participation rates.^39,41 In Turk et al., 20 percent of care providers in the intervention arm stated they had not been given a personal health profile (PHP) at followup, only 18 percent of patients and 39 percent of care providers who said they received the PHP reported using the tool, and less than a third of care providers who said they received the PHP reported taking the tool to a primary care visit.

The four previously described studies prompt physician action or physician-patient interaction at the point of care. The last study in this category prompts patients to interact with the health care system.³⁷ Women with mobility disabilities participated in a two-hour small group (four women) workshop led by women who also had mobility disabilities. The workshop covered not only general screening benefits, but also ways to overcome physical barriers to screenings and how to communicate with physicians to facilitate screenings.³⁷ Participants in the intervention arms also received structured phone calls for 6 months. Compared with a written information control, women in the intervention group self-reported receiving significantly more pap testing; there was no significant difference in mammography rates between the attention and control groups. The two major weaknesses of this study are the lack of attention control and the self-reported screening outcomes.

Interventions Improving Access to Care

The four trials on access to care are part of a much larger literature on virtual care for people with disabilities. Many virtual access articles were excluded during title/abstract screening because the intervention occurred outside the formal health care system; that is, the study did not involve health care practitioners delivering health care virtually, or the focus of the study was not creating access based on the nature of the disability. For example, Morland conducted two trials of virtual care for veterans with post-traumatic stress disorder (PTSD).^47,48 However, the remote care required people in remote areas to travel to a central location for group therapy. This intervention seemed more focused on creating access due to workforce shortages than creating a way to deliver care that would feel safer and less stigmatizing.

The inclusion status of the final set was influenced by article framing. Articles were framed as follows: “Despite the findings supporting the Managing Fatigue program [for people with multiple sclerosis], its major limitation to date has been its inaccessibility to individuals who cannot travel to the community sites where the program is offered;” (p.1131)⁴³ “The Internet might provide an alternative information and treatment opportunity for people who avoid care because of concerns about the stigma of receiving mental health treatment;” (p.73)⁴² “For individuals with RA, travel may be difficult due to pain or functional limitations;” (p. 1578)⁴⁴ and “Telephone counseling was selected as the intervention delivery mode because it is an effective behavior change approach that overcomes barriers to participation such as distance, accessibility, and limited transportation, barriers that may be even more prominent among people with MS.”(p.90).⁴⁹

Table 8 summarizes the included studies providing virtual access to care.

Table 8

Summary of interventions providing virtual access to care.

While all of the studies in this group used a virtual access to care intervention, each study used different populations, intervention characteristics, and targeted outcomes. Use of waitlist controls was the only element common across studies.

The primary outcome for the counseling intervention to decrease major depression by increasing physical activity in people with MS was a reduction in depression severity, as measured by the Hamilton Depression Rating Scale (HAM-D).⁴⁹ A clinically significant response to treatment was defined as a 50 percent decrease in the total HAM-D score. There was no clinically significant difference in depression severity between the treatment control groups. However, treating HAM-D scores as continuous, there was a significant decrease in depression severity among the treated, negative affect, and number of people meeting the criteria for major depressive disorder or dysthymia among the treated.⁴⁹

The primary outcomes for the teleconference intervention aimed at managing fatigue in individuals with MS include: fatigue severity (measured using the Fatigue Impact Scale [FIS], fatigue impact (measured using the Fatigue Severity Scale [FSS]), and health-related quality of life (measured using the SF-36 Quality of Life Scale).⁴³ An intent to treat analysis found significant effects of the intervention on all three subscales of the fatigue impact severity measure and the role physical subscale of the SF-36; fatigue severity, self-efficacy, and the other seven domains of the SF-36 did not differ significantly.⁴³

Primary outcomes of the therapeutic writing intervention for PTSD include measures of: symptoms of posttraumatic stress (Impact of Event Scale, IES-R), depression and anxiety (SCL-90), self-reported physical and psychological function (SF-12), and patient and therapist agreement (Working Alliance Inventory [WAI]).^42,50 Participants in the intervention arm showed significant improvements over time on all measures except physical health, compared with the waitlist control.⁵⁰

Rheumatoid arthritis (RA) online outcomes included rheumatic disease specific self-report of health status and well-being (Arthritis Impact Measurement Scales 2, AIMS2), an arthritis specific self-efficacy measure (Arthritis Self-Efficacy Scale [ASES]), depression (Center for Epidemiologic Studies Depression Scale, CES-D), quality of life (QLS-15), measure of joint and pain tenderness (Rapid Assessment of Disease Activity in Rheumatology, RAPID), a measure of social connectedness (Social Provisions Scale, SPS), and a measure of loneliness (University of California, Los Angeles Loneliness Scale, version 3, LS-3).⁴⁴ Immediately after interventions, significant gains in self-efficacy and quality of life were observed; these gains were maintained for 9 months post-intervention.⁴⁴

Overview

Despite the difficulty in locating studies that fit a cultural competence for disabilities concept due to the lack of commonly used indexing terms and the minimal application of the cultural competence concept to disability research, we identified 25 intervention studies. These 25 studies map to only a small portion of the many and diverse groups that comprise the overall set of people with disabilities.

The most traditional cultural competence intervention, provider training, was also the most common type of study found in this review. This literature was dominated by interventions aimed at reducing stigma and changing provider attitudes toward people with mental illness. Only four studies attempted to change provider attitudes and beliefs about people with physical or intellectual disabilities, two of which attempted to do so through having students participate in a disability simulation (an approach with the potential to confound stereotypic thinking). Contact-based interventions, in which providers interacted with a person with a disability, were the most common intervention type; two contact-based interventions resulted in more negative attitudes among some participants.^23,45 Twelve of the 16 included trainings or curricula studies developed for students, not working professionals. Many of the available studies were not designed to capture how well initial knowledge gains or changes in attitudes are sustained over time. Of those that did have a longitudinal design, one study found changes were not sustained for 6 months,²⁹ two studies found changes were sustained for 2 and 6 months respectively.^27,33 The effectiveness of these interventions depends on students applying their pre-service training to their work in clinical settings.

We considered two other types of interventions in this review: the patient or physician prompt at the point of care; and virtual care that creates disability sensitive access. Lennox and colleagues are pioneering low-tech, seemingly effective ways to prompt universal screenings for areas of known disparities for persons with intellectual disabilities. While promising, we did not find additional high quality trials to support this work. If there is little evidence for upstream interventions, such as provider trainings,⁵¹ the research community may want to also focus on downstream, point of care interventions. Interventions that prompt the physician at the point of care work hand-in-hand with interventions that empower the person with disabilities to talk to their doctor.³⁷

We also broadened our definition of cultural competence to include virtual interventions that addressed disability related barriers to access including difficulty traveling and not wanting to get care because of stigma. These studies have the most patient-centered outcomes. Just as focusing on the point of care does not mean giving up on changing provider attitudes and beliefs, creating a parallel system for care does not preclude making the mainstream system more disability competent.

Research Directions

The risk of bias for most of the included studies was high (Appendix D). The major weaknesses of this literature are the lack of attention controls, lack of appropriate comparators, attrition, and social desirability bias of pre- post attitude scales. Michaels et al. is the only included study to try to address social desirability among respondents.³⁴ More well-designed studies are needed to test the effectiveness of disability competence interventions, particularly for nonmental health related disabilities. We need to be looking for unintended effects of trainings, and the effect of trainings and one time interventions on patient-centered outcomes. Without attending to methodological concerns, it will remain difficult to answer whether such interventions improve care and reduce health disparities.

Cultural competence is not a one-size-fits-all concept across populations that experience health disparities. Much of the team's work during this review was spent defining the concept of cultural competence for the disability population. The work of developing definitions for cultural competence as well as effective solutions for improving providers' knowledge and training in the health needs of people with disabilities should involve people with disabilities. While community-based participatory research with racial and ethnic groups has a fairly strong track record, much could be done to bring the perspective of people with disabilities into the research process. Future research on disability-related health care disparities and interventions to address them should target dimensions most important to people with disabilities and include more patient-centered outcomes. Including people with disabilities in research conceptualization and design is critical to identifying more effective solutions and producing evidence that could be understood and used by various stakeholders including people with disabilities.

Many disability groups and subgroups are not represented in this literature. Many subgroups exist within the disability populations with multiple perspectives, interests, and challenges. These differences can be further complicated by factors not directly attributable to disability, such as rural or urban location, poverty status, or racial and ethnic differences. Interventions targeting the intersection of populations of interest (such as race and disability) were also not well researched for this population. Researchers have begun to document health disparities at the intersection of disability and race/ethnicity.^52-56 There is also movement to align disparities research across race/ethnic and disability populations.^7,55,57 However there is not a sufficient evidence base to conclude whether interventions used to promote racial and ethnic provider cultural competence will produce reductions in disparities when used to promote provider cultural competence for people with disabilities in health care contexts. Carefully designed studies conducted for race/ethnic and disability populations, as well as their intersection, are needed. Intervention types appear to be tied to disability types in the literature. Work is needed to conceptualize cultural competence to address inclusion of the many populations and interventions under the diverse disability umbrella.

Limitations

This review is limited by the difficulty of locating literature using either MeSH® terms or natural language keywords. This difficulty is exacerbated by the new extension of the concept of cultural competence to disability cultures. This, in turn, may be influenced by a lack of consensus among disability communities about whether disability is a “culture.”

While care and attention was dedicated to defining the scope boundaries for this review, they are necessarily arbitrary; no clear lines of demarcation can be easily drawn to separate patient-centered care, health literacy, or other quality improvements from cultural competence. Interventions that focused solely on changing the patient (e.g., patient education and health promotion) were excluded because while they targeted a reduction in a health disparity experienced, they did not require change on the part of the physician or the health care system. We also excluded wellness and secondary disease prevention trials that did not target the formal health care delivery system or its providers. Those studies are, however, relevant to the larger discussion of reducing health disparities in this population.^58-74

Only eight of the 25 included studies were conducted in the United States. This raises questions regarding transferability of the included studies to the U.S. health care system.

Conclusions

The results of the search show a patchy literature set that highlights the intrapopulation diversity subsumed under the disability umbrella. The literature also fails to recognize the intersections of disability populations with other populations experiencing healthcare disparities. The broader concept of diversity competence may be more appropriate for many people at these intersections. Further discussion of population intersectionality and alternative constructions of cultural competence that address structural inequities can be found in Chapter 5. None of the included studies measured the effect of cultural competence interventions on health care disparities. The medium or high risk of bias of the included studies, the heterogeneity of populations, and the lack of measurement consensus prohibited pooling estimates or commenting about efficacy in a meaningful or responsible way.