As curricula to improve medical students' attitudes toward people with disabilities are developed, instruments are needed to guide the process and evaluate effectiveness. The authors developed an instrument to measure medical students' attitudes toward people with disabilities. A pilot instrument with 30 items in four sections was administered to 342 medical students. Internal consistency reliability and factor analysis were conducted. The Cronbach's alpha coefficient was 0.857, indicating very good internal consistency. Five components were identified: comfort interacting with people with disabilities, working with people with disabilities in a clinical setting, negative impressions of self-concepts of people with disabilities, positive impressions of self-concepts of people with disabilities, and conditional comfort with people with disabilities. The instrument appears to have good psychometric properties and requires further validation.
According to a U.S. Census Bureau report, about 54.4 million people (18.7% of the U.S. population) reported some level of disability in 2005, and 35 million of them (12% of the population) were classified as having a severe disability (U.S. Department of Commerce, Economics and Statistics Administration & U.S. Census Bureau, 2008). Healthy People 2010 identifies people with disabilities as a vulnerable and at-risk population, subject to health care disparities including increased rates of secondary conditions such as heart disease, lack of access to health care services, and lower rates of screening and preventive care services (U.S. Department of Health and Human Services, 2000).
Disability categories are vast and include congenital physical disabilities (e.g., cerebral palsy, spinal bifida), developmental disorders (e.g., autism, mental retardation), sensory disabilities (e.g., deafness, blindness), chronic disease (e.g., cancer, diabetes), and mental illness (e.g., schizophrenia, dementia). The World Health Organization (n.d.) recognizes that disability is a complex phenomenon, reflecting an interaction between a person's body and features of the society in which he or she lives. The Americans with Disabilities Act of 1990 defines disability as physical or mental impairment that substantially limits one or more of the major life activities of the individual (U.S. Department of Justice, 2009). Healthy People 2010 identifies people with disabilities as persons having an activity limitation, who use assistance, or who perceive themselves as having a disability. It stresses that, for the purposes of improving care of patients with disabilities, the similarity among people with disabilities is as important as or more important than the differences among clinical diagnostic groups (U.S. Department of Health and Human Services, 2000).
Barriers for people with disabilities to accessing health care services include physical factors (inaccessibility of facilities and equipment to people with disabilities); deficiencies in the knowledge, attitudes, and skills of health care providers; and negative attitudes and behaviors of health care providers (Drainoni et al., 2006; Field, Jette, & Institute of Medicine [U.S.] Committee on Disability in America: A New Look, 2007; Morrison, George, & Mosqueda, 2008). Negative attitudes held by health professionals may negatively influence the quality of health care, resulting in withholding treatment, giving inferior treatment, and neglecting general and preventive care (Jackson, 2007; Paris, 1993). Negative attitudes may also compound adverse outcomes and limit services available to people with disabilities (Tervo, Palmer, & Redinius, 2004). For example, a doctor may not perform a pelvic exam on a female patient who has a neurologic injury, assuming that she is not sexually active. Many patients with disabilities have reported finding their doctors insensitive or patronizing (Byron, Cockshott, Brownett, & Ramkalawan, 2005). Participants in a focus group study reported that the attitudes of health care professionals were poor and that the clinicians focused on the disability rather than the patient's chief complaint (Morrison et al., 2008). It has been suggested that clinicians' negative attitudes could be due to a lack of training or inexperience with people with disabilities (Aulagnier et al., 2005). Many primary care physicians have reported being uncomfortable working with people with disabilities (Larson McNeal, Carrothers, & Premo, 2002; Shah, Norlin, Logsdon, & Samson-Fang, 2005).
Students in the health care professions often harbor negative attitudes toward patients with disabilities (Au & Man, 2006; Matziou et al., 2009; Tervo, Azuma, Palmer, & Redinius, 2002; Tervo et al., 2004) and are often uncomfortable interacting with them (Jain, 2006; Shah et al., 2005). However, there is evidence that prior experience interacting with people with disabilities as well as early and frequent encounters with patients with disabilities during training improve students' attitudes (Brown, Graham, Richeson, Wu, & McDermott, 2010; Jackson, 2007; Larson McNeal et al., 2002; McKenna, Scholtes, Fleming, & Gilbert, 2001; Messmer, Conway, Giltinan, & Stroh, 1990; Tervo et al., 2004; Thistlethwaite & Ewart, 2003). Despite this, caring for patients with disability is rarely addressed in medical school curricula (Byron et al., 2005; Crotty, Finucane, & Ahern, 2000; Jackson, 2007; Kahtan, Inman, Haines, & Holland, 1994; Martin, Rowell, Reid, Marks, & Reddihough, 2005). Following calls on the local (Symons, McGuigan, & Akl, 2009), national (Field et al., 2007; Kirschner & Curry, 2009; U.S. Department of Health and Human Services, 2000, 2005), and international (World Health Organization, n.d.) levels, several medical schools—including our own—are currently involved in implementing curricula to improve students' knowledge, attitude, and skills in this area (Byron et al., 2005; Crotty et al., 2000; Graham, Brown, Zhen, & McDermott, 2009; Jackson, 2007; Kahtan et al., 1994; Kirschner, Breslin, Iezzoni, & Sandel, 2009; Martin et al., 2005; Symons et al., 2009). If these curricula are to be robust, there is a need for valid and reliable instruments to guide their development. This includes establishing baseline measures to evaluate the effectiveness of new curricula by documenting changes after implementation.
We conducted a systematic review (Lam et al., 2010) of the medical literature to identify validated instruments to measure the attitudes of health care students and professionals toward people with disability. An instrument was needed to evaluate the implementation of a curriculum that instructs students about the care of people with disabilities. We identified seven validated instruments, but none was designed specifically for medical students. Three were designed for health care students or professionals (Dunn, Umlauf, & Mermis, 1992; Lee & Sonis, 1983; Messmer et al., 1990; Rice, 1979), and four were designed for the general population (Antonak, 1981; Gething, 1992; Yuker, Block, & Younng, 1970; Yuker & Hurley, 1987). Also, the most recent instrument was developed in the early 1990s (Gething, 1992), and many used terminology that is not relevant or socially accepted today. The objective of our study was to develop an instrument to measure medical students' attitudes toward people with disabilities.
We identified potential items for the instrument through a systematic review of previously published instruments (Lam et al., 2010). The instrument was reviewed and refined by medical educators at our school and administrators from three major community-based agencies specializing in health care and social services for people with disabilities. In addition, a patient advocacy group and a family council reviewed the instrument and suggested modifications based on their own experience with the health care system. Some items were changed to reflect currently acceptable language in the field, and others were changed to capture well-established attitudinal concerns toward people with disability. Scenarios were added to the instrument to make it specific to medical students. Construct validity of the pilot instrument was then established through a cycle of instrument item reviews and revisions by experts in the field of treatment of people with disabilities.
The resulting instrument consisted of 30 items organized in four sections (see the Appendix):
Two were demographic items on gender and age.
Two were factual items, asking participants if they had any professional (i.e., volunteering at a group home, teaching) or personal experience (i.e., friend or relative with a disability) with people with disabilities.
Eighteen were opinion questions asking participants to indicate on a four-point scale (strongly disagree, disagree, agree, strongly agree) their beliefs about people with disabilities and/or how comfortable they would be with people with disabilities in specific encounters. This section included items such as “People with disabilities are as happy as people without disabilities” and “If I were visited by a person who is blind, I would be comfortable helping him or her navigate the environment.”
The last section consisted of descriptions of two clinical scenarios. The scenarios were identical except that in one the patient does not have an apparent disability and in the other the patient is in a wheelchair and appears to have spasticity and speech impairment. Each scenario was followed by four questions. The first question was factual, asking whether the participant had ever encountered such a clinical situation. The remaining three questions in each scenario were about the expected level of comfort in handling such a situation. These six items used the same four-point scale used in the previous section of the instrument.
The primary purpose of the instrument is to evaluate curricula aimed at improving student attitude toward people with a disability by measuring that attitude both before and after classroom instruction and clinical experiences. Therefore, the sample used to collect data for psychometric analysis included students across the years of medical education. The final sample of 342 students included 246 (71.9%) first-year students, 79 (23.1%) second-year students, and 17 (5.0%) third-year students from two medical schools in New York State. Of the 338 students who indicated their gender, 52.0% (178) were female. The mean age of the participants was 24.2 years. Of the 331 students who responded to the questions about professional and personal experience with people with disabilities, 51.7% (171) had professional experience, and 44.1% (146) reported personal experience. Ninety-three (28.1%) of the students reported both professional and personal experience with people with disabilities.
Psychometric analysis was conducted using responses to the 24 scale items in the instrument—the 18 opinion questions and the three questions about comfort in handling the situation in each of the two scenarios. Internal consistency reliability analysis and factor analysis using the principal components method of extraction were applied to the data collected to establish internal consistency of the scale and to identify a factor structure. Six items were reverse-scored prior to analysis because higher ratings on those items indicated a more negative attitude toward people with disability, while higher ratings on all other items indicated a more positive attitude.
We determined the suitability of the data for factor analysis through examination of the sample size and the correlation of items. First, the sample size of 342 participants met the general rules of at least 150 cases and/or a ratio of at least five cases for each of 24 scale items included in the analysis. Second, although correlations among items were very low (generally less than 0.3), indicating that factor analysis may not be appropriate, the Kaiser-Meyer-Olkin measure of sampling adequacy was 0.805, exceeding the suggested minimum of 0.6, and the Bartlett's test of sphericity was significant (p = .000), indicating that these data were appropriate for running factor analysis. SPSS version 16.0 (SPSS Inc., 2007) was used for all analyses.
Internal Consistency Reliability
Internal consistency reliability analysis resulted in a Cronbach's alpha coefficient of 0.857 for this scale, with no notable increase possible through deletion of any items. This indicates very good internal consistency reliability for this instrument with this sample. Four items had low (less than 0.3) corrected item-total correlation values. While this may indicate that these items were measuring something different from the scale as a whole, the principal component analysis (PCA) provided more information about the suitability of these items.
Factor analysis using the principal components method of extraction was run using oblique rotation. The correlations among components were low (0.003 to 0.243), indicating that it was appropriate to use orthogonal rotation instead. PCA using orthogonal rotation revealed six components with eigenvalues of 1.0 or more. Horn's parallel analysis was run and revealed that four eigenvalues from the sample data exceeded those from a randomly generated data set, indicating that four components should be considered.
While the results of the parallel analysis are generally found to be most accurate, it is also important to find an accurate solution based on there being a meaningful group of survey items in each component. To examine this, different numbers of components were explored in the second step of the PCA factor rotation. A minimum item loading of 0.4 was specified in this step, and a five-component solution resulted in the most meaningful item groupings (see Table 1). The components were
Comfort interacting with people with disabilities (Items 4, 6, 7, 9, 12, 14, 18, and 19 in the Appendix). Example item: “I would be comfortable interacting with a person with an intellectual disability who was in the community on his or her own (i.e., without staff members or caretakers).”
Working with people with disabilities in a clinical setting (Items 22, 23, 24, 26, 27, and 28 in the Appendix). Example item: “I would be comfortable performing a physical exam on the patient.”
Negative impressions of self-concepts of people with disabilities (Items 3, 8, and 11 in the Appendix). Example item: “Most people with disabilities feel sorry for themselves.”
Positive impressions of self-concepts of people with disabilities (Items 5, 13, and 17 in the Appendix). Example item: “Most people with disabilities are not ashamed of their disability.”
Conditional comfort with people with disabilities (Items 10 and 15 in the Appendix). Example item: “I am more comfortable around people with intellectual disabilities when they have someone who is not disabled to help them.”
During the PCA analysis, one item did not load on any component with the minimum value of 0.4 and was removed from the analysis (Item 16 in the Appendix). Another item loaded highly on two components and was also removed (Item 20 in the Appendix). An examination of the two items removed revealed that they may have been ambiguous, resulting in a lack of a pattern in participant responses.
Among the reasons cited for health care disparities for people with disabilities is the lack of comfort experienced by caregivers in treating people with disabilities as well as attitudinal misconceptions of providers toward this population (Drainoni et al., 2006; Field et al., 2007; Jackson, 2007; Morrison et al., 2008). These attitudes may compromise care and adversely affect outcomes (Antonak & Livneh, 1988, 2000,; Dunn et al., 1992; Mitchell, Hayes, Gordon, & Wallis, 1984; Paris, 1993; Yuker, 1986). A “positive attitude” toward people with disabilities has been described in the literature as the belief that people with disabilities can be productive members of society, make decisions regarding their own self-interest, and lead a normal life (Tervo et al., 2004). Caregivers who possess a positive attitude toward people with disabilities can create conditions to help individual patients achieve their capacity and contribute to the community (Tervo et al., 2004). Attitudes may exist in several realms: affective (i.e., emotions), cognitive (i.e., thoughts and beliefs), and behavioral (i.e., actions [Rajecki, 1990]). Our institution has instituted a curriculum throughout the four years of medical school to help medical students gain experience working with people with disabilities (Symons et al., 2009). Our hope is that early exposure to people with disabilities will help to increase the comfort level of students as well as to foster attitudes conducive to the provision of quality care. The development of our attitudinal assessment tool will help to guide and evaluate this process. Documenting changes in attitude and comfort level before and after the implementation of a particular element of the curriculum will help assess its efficacy in achieving the overall goals of the curriculum.
Several tools have historically been used to assess attitudes toward people with disabilities. Attitudes Toward Disabled People (Yuker et al., 1970) is the most widely used and tested instrument (Antonak, 1980; Antonak & Livneh, 1988, 2000; Block, 1974; Chan, Lee, Yuen, & Chan, 2002; Lee, Paterson, & Chan, 1994; Tervo et al., 2004; Yuker, 1986; Yuker, Block, & Campbell, 1960; Yuker, Block, Younng, & Human Resources Center, 1966). Its 30 Likert-like statements seek to measure the extent to which the respondent perceives persons with disabilities to be similar to or different from the general population. Although the Attitudes Toward Disabled People scale is the most widely applied, concerns have been raised regarding its unidimensional structure and outdated items (Antonak, 1980; Antonak & Livneh, 1988; Livneh, 1985; ten Klooster, Dannenberg, Taal, Burger, & Rasker, 2009). The Scale of Attitudes Toward Disabled Persons (Antonak, 1981) is another measure of attitudes of the general population toward people with disabilities in general, and it is also often applied to health care professionals and students. Its items examine three factors related to the general population's attitude toward people with disabilities: optimism/human rights, behavioral misconceptions, and pessimism/hopelessness (Antonak, 1982). The Interactions with Disabled Persons Scale (Gething, 1992) examines levels of discomfort in social interactions. The Multidimensional Attitudes Scale (Findler, Vilchinsky, & Werner, 2007) looks at three dimensions of attitude (affective, cognitive, and behavioral) among the general population. Although these scales have been applied to health care workers and students (Lee & Sonis, 1983; Matziou et al., 2009; ten Klooster et al., 2009; Tervo et al., 2002; Tervo et al., 2004), none has been specifically designed for and validated with medical students.
Our instrument addresses five attitudinal dimensions of medical students toward people with disabilities (as described above in our factor analysis). Notably, the inclusion of the two patient-care scenarios provides an opportunity for respondents to project their own emotions, thoughts, and behaviors onto the given clinical situation. The inclusion of this element in the instrument also creates the option of interchanging different scenarios to adapt the tool for other types of health care professionals and for specific areas of disability.
Two demographic items are included in our instrument: age and gender. When we administered the survey for the purpose of psychometric analysis, we excluded an item addressing the student's year in medical school because we were aware of which class completed the instrument. However, when using the tool in other settings, educators and researchers may include this and other demographic questions.
A limitation of our instrument is that it relies on self-reported attitudes. Roby and colleagues (Robey, Beckley, & Kirschner, 2006) note that respondents may not necessarily have the ability or motivation to record their beliefs about people with disabilities accurately. Though methods such as the Implicit Association Test (Greenwald & Banaji, 1995), which measure unconscious implicit associations made regarding people with disabilities, tap into attitudinal dimensions not measured by self-reported surveys, these methods are relatively time- and resource-intensive and would be difficult to administer to an entire medical school class. Roby and colleagues did find some evidence of a relationship between implicit and explicit (i.e., self-reported) attitudes (Robey et al., 2006).
In conclusion, we have developed an instrument to measure medical students' attitudes toward people with disabilities. The instrument appears to have good psychometric properties and requires further validation. The instrument will help identify potential concerns in students' attitudes and thus establish the need for a curricular intervention to address these attitudes. Educating medical students about caring for patients with disabilities is critical to assuring quality health care for this population (Kirschner & Curry, 2009). Although the ultimate measure of our curriculum's efficacy will be documentation of increased patient satisfaction and outcome, our tool will be instrumental in documenting changes in student attitude, which will serve as the basis for changes in physician practice.
Editor-in-Charge: Stephanie Dean
Andrew B. Symons, University at Buffalo; Reva Fish, State University of New York College at Buffalo; Denise McGuigan (e-mail: firstname.lastname@example.org), University at Buffalo School of Medicine and Biomedical Sciences, Family Medicine, 202 Farber Hall, Buffalo, NY 14214, USA; Jeffery Fox, Canisius College; and Elie A. Akl, email@example.com.