FLAMOS 1
Running head: FOREIGN LANGUAGE ANXIETY IN A MEDICAL OFFICE SCALE
The Foreign Language Anxiety in a Medical Office Scale (FLAMOS): Developing and Validating a Measurement Tool for Spanish-Speaking Individuals
Abstract
Communication research has been hindered by a lack of validated measures for Latino populations. To develop and validate a foreign language anxiety in a medical office scale (FLAMOS), a survey of low income, primarily Spanish-speaking Latinos (N = 100) was conducted. The scale demonstrated convergent, divergent, and predictive validity, high reliability (α = .92), and a uni-dimensional construct. FLAMOS provides a validated measure for researchers and may help to explain Latino healthcare communication barriers.
Keywords: Latino, health, communication apprehension, Spanish, ESL, foreign language anxiety
The Foreign Language Anxiety in a Medical Office Scale (FLAMOS): Developing and Validating a Measurement Tool for Spanish-Speaking Individuals
The U.S. Latino population has been identified as an underserved community with lower access to quality health care (American Medical Association, 2004). Latinos are the fastest growing population in the United States; the Census Bureau (2006) has speculated that as much as 20% of the U.S. population will be Latino by 2035. In recognition of this population shift, Healthy People 2010 noted that more attention needed to be paid to addressing cultural and language barriers within a health context, so as to lessen the healthcare disparity gap between underserved populations and populations that have resources to quality care (U.S. Department of Health and Human Services, 2001). As such, healthcare providers should be prepared to address barriers in health communication that are unique to the growing Latino population.
Language issues are a primary concern of healthcare providers when treating non-English speaking populations in the United States (Flores, 2005). Language barriers can create a variety of problems in a medical context, particularly when the physician and patient do not share a common language (Clark, Sleath, & Rubin, 2004). Communicating across languages can compromise patient-physician communication and the quality of information that is exchanged (Flores, 2005; Rivadeneyra, Elderkin-Thompson, Silver, & Waitzkin, 2000).
Patient fears about communicating can also influence health communication processes and the patient’s willingness to seek out or provide health information (Booth-Butterfield, Chory, & Beynon, 1997). Specifically, individuals may be anxious about communicating in another language, and thus be less willing to provide information or ask questions, less able to adequately describe themselves, and less able to accurately interpret or translate information (Ganschow & Sparks, 1996; MacIntrye & Gardner, 1994). To date, however, the potential communication effects for when individuals have high foreign language communication anxiety have only been studied in populations of students learning a foreign language (e.g., Horwitz & Young, 1991). If this anxiety transfers to individuals in language discordant medical interactions, anxiety may inhibit patients' ability to communicate in the situation, exacerbating existing language barriers.
In order to properly explicate any construct, such as foreign language anxiety in medical office settings, and study its potential causes, effects, and outcomes, a measurement tool needs to be created to adequately operationalize the construct (Chaffee, 1991). Communication research has been hindered by a lack of validated measures (Chaffee, 1991; Rubin, Palmgreen, & Sypher, 2004). This is particularly problematic to the advancement and testing of communication theory, as scientific study tends to mature in direct relation to the sophistication of available measurement instruments (DeVellis, 2003; Schmidt & Hunter, 1999).
The present study seeks to develop and validate the foreign language anxiety in medical office scale (FLAMOS). Three research areas informed the development of the scale: research regarding language barriers for Latinos in receiving quality healthcare (e.g., Flores, 2005), studies of communication anxiety and its effects (e.g., Booth-Butterfield et al., 1997), and work examining foreign language communication anxiety in students learning a different language (e.g., MacIntrye & Gardner, 1989). FLAMOS presents an intersection of these areas of research by providing a tool to identify and improve outcomes of individuals with high anxiety about language discordant communication in medical office settings.
Healthcare Language Barriers for Latinos in the United States
Half of Latinos living in the U.S. who speak Spanish at home report having problems communicating in English (U.S. Census Bureau, 2008). Kim’s (2001) theory of cross-cultural accommodation posits that communicating in the language of the host country is central to acculturation processes. Individuals with low English proficiency (LEP) can have problems negotiating the primarily English healthcare system in the United States (Clark et al., 2004). Although professionally trained bilingual healthcare staff or interpreters are sometimes available (Flores, 2005), the majority of healthcare providers do not have access to interpretation services, particularly those in rural areas, (Kuo, O’Connor, Flores, & Minkovitz, 2007).
Having to navigate a predominantly English healthcare system has been shown to produce a number of undesirable outcomes for Latino patients that have low English health literacy (Clark et al., 2004). Specifically, LEP may be directly related to health status (Hoffman-Goetz, Meissner, & Thomson, 2009). Difficulties created by language barriers can be far reaching, ranging from issues with access to healthcare (DuBard & Gizlice, 2008), issues with non-professional interpreters such as children (Flores, 2005; Morales & Hanson, 2005), a lack of adherence or understanding of treatment (Westberg & Sorensen, 2003), problems with printed medication instructions (Bradshaw, Tomany-Korman, & Flores, 2007), and basic patient-physician communication (Clark et al., 2004). LEP populations are less likely to participate in preventive care, have poorer patient satisfaction, and less knowledge about the availability of services and what benefits they provide (Dilworth, Mott, & Young, 2009).
Language can be a serious barrier to adequate health communication with LEP Latino patients. Specifically, predominantly Spanish-speaking Latinos have a harder time finding and interpreting health information, and thus are less likely to seek important health information, such as knowledge about cancer and cancer prevention (Davis, Diaz-Mendez, & Garcia, 2009; Vanderpool, Kornfeld, Rutten, & Squiers, 2009). Primarily Spanish-speaking patients are less likely to ask questions when they are interacting with an English-speaking physician and less likely to receive information from the physician (Schouten & Meeuwesen, 2006).
One aspect of interpersonal interaction that likely influences the quality of patient-provider communication is communication apprehension(Booth-Butterfield et al., 1997). Communication apprehension in a medical setting is a commonly experienced phenomenon among patients (Bowden & Burstein, 1979). Patients often become anxious about communicating health issues with a stranger who is very busy (Foley & Sharf, 1981). Communication apprehension can influence how patients and physicians interact and the patient’s willingness to seek information (Booth-Butterfield et al., 1997).
Communication Anxiety/Apprehension
Communication apprehension (CA) is a commonly studied phenomenon in communication research, and scholarship has focused on theorizing about CA, adequately measuring CA, examining specific situations in which CA becomes salient, and examining outcomes influenced by CA. Research has focused on identifying and theorizing about communication apprehension, defined as the "level of fear or anxiety associated with either real or anticipated communication with another person or persons" (McCroskey, 1977, p. 78).Alternatively, the phrase “communication anxiety” is used as a synonym to communication apprehension (see Booth-Butterfield & Gould, 1986). For the purpose of this paper, the terms communication anxiety and communication apprehension are used interchangeably.
CA predicts various communication behaviors and other outcomes when apprehension is very high or very low (Beatty, 1987). Communication anxious individuals forced into communication situations often withdraw by communicating less verbally (McCroskey, 1976), and by attempting to shorten encounters (Lazarus & Averill, 1972). Wheeless (1975) notes that CAmay stem from having to communicate in a particular situation, but may also arise when receiving information or being the communication recipient. Although individuals are less apprehensive when they are the recipient rather than the source, high communication recipient anxiety is still present in some individuals (Wheeless, 1975). One particular area of study has focused on CA when individuals have to communicate in a language that is not their primary or native language (McCroskey, Fayer, & Richmond, 1985; Horwitz & Young, 1991). Several studies have examined this foreign language communication anxiety (e.g., MacIntyre & Gardner, 1991).
Foreign Language Communication Anxiety
Individuals are often more anxious about communicating in a secondary or foreign language than they are about communicating in their primary language (McCroskey, Fayer et al., 1985; Richmond, McCroskey, McCroskey, & Fayer, 2008). The concept of being apprehensive about communicating in a language that is not the speaker’s native language, or foreign language communication anxiety(FLCA), has been primarily studied in students learning a foreign language in school, or in exchange students visiting a country that speaks a language different than their native language (e.g., Horwitz & Young, 1991). FLCA is distinct from other types of communication anxiety (MacIntyre & Gardner, 1989) and from an individual’s proficiency with a foreign language (MacIntyre, Noels, & Clément, 1997).
MacIntyre and Gardner (1989; 1994) provide a model of how FLCA occurs at three levels. Students can be anxious when the foreign language is encountered (input), when processing the foreign language (processing), and when using the foreign language (output). Their model also posits how FLCA interacts with general trait-based communication apprehension and language performance. The more that students are apprehensive about communicating in the language, the worse they will perform in using the language, causing greater anxiety both in foreign language communication and general communication. MacIntrye and Gardner (1989) posit that this cycle will solidify a context-based anxiety (for example, French class anxiety) and that this cyclical process will continue, strengthening and maintaining the anxiety felt when interacting in the specific context (e.g., when communicating in French class).
A number of studies have shown that high FLCA can lead to negative outcomes for students. Students with higher FLCA have problems in the processing stage, as they translate less accurately, comprehend less, and guess at unknown words and phrases less often than their non-anxious counterparts (Ganschow & Sparks, 1996; MacIntrye & Gardner, 1994). The discrepancy in translation accuracy can disappear if highly anxious students put more effort into their studies (MacIntrye & Gardner, 1994). In terms of output, higher FLCA can lead to a lower quality language performance than that of students at similar proficiency levels who are not anxious communicating in the language (MacIntrye & Gardner, 1989). Students struggled with vocabulary and giving quality descriptions of themselves when they were high in FLCA, potentially because of their reduced ability to remember vocabulary words (MacIntrye & Gardner, 1994). Students high in FLCA had lower final course grades than students who were moderately anxious or students who were low in anxiety (Ganschow & Sparks, 1996). Specifically, FLCA was found to have a greater influence on language performance as students advanced to higher-level language courses (Saito & Samimy, 1996).
Foreign Language Anxiety in a Medical Context
Research on FLCA has been conducted extensively on students in classrooms, but has not examined the effects of FLCA in other contexts with non-student participants. The outcomes influenced by FLCA in a classroom could have far-reaching implications if they are replicated in patients seeking medical care offered in their secondary language (e.g., monolingual or primarily Spanish-speaking Latinos living in the United States). If patients are high in FLCA and less able to accurately interpret, translate, and comprehend medical information, they will be more likely to have problems pursuing preventive health practices, could receive lower quality healthcare, and have difficulty understanding and adhering to medical treatment. Furthermore, patients high in FLCA might be less able to give quality self-descriptions of health problems, leading to misdiagnoses, mistreatment, or a lack of diagnosis.
Health communication research focusing on issues in Latino populations related to quality delivery of health services indicates that FLCA, and adverse outcomes related to FLCA, might be present. Research has identified that language issues are problematic in a health context, but Dilworth and colleagues (2009) noted that research examining how to address these issuesis needed. Spanish-speaking patients received the poorest quality healthcare when paired with English-speaking physicians, regardless of the patients’ health literacy level (Sudore, Landefeld, Perez-Stable, Bibbins-Domingo, Williams, & Schillinger, 2009). Health information seeking is also affected by language difficulties and anxieties. For example, eighty percent of Latinos do not search for cancer information, and those that do report problems and frustration with understanding the information (Vanderpool et al., 2009). Specifically, Latinos report a fear of discussing cancer in English (Davis et al., 2009). General communication apprehension is shown to negatively influence “levels of question-asking, understanding, and length of contacting,” as well as information seeking and positive descriptions of patient-physician interactions (Booth-Butterfield et al., 1997, p. 246). The fact that Latinos report these outcomes specifically within an English-language setting suggests that FLCA may influence information seeking and patient-physician communication. This is problematic as individuals that report frustrations or difficulty finding cancer information are less likely to report being in excellent or good health condition (Hoffman-Goetz et al., 2009). Additionally, physicians are less likely to give as much medical information to patients high in communication anxiety compared to patients low in communication anxiety (Graugaard, Eide, & Finset, 2003). Patients with high FLCA may experience poor healthcare treatment because of their unwillingness and inability to produce high quality self-descriptions and because of the physician’s tendency to provide less information.
The present study focused on the development of a scale that can be used to measure FLCA in a medical setting. Self-report measures are an effective way to measure communication anxiety (Wheeless, 1975). Since this anxiety is a type of fear created by the way that individuals cognitively process situations, the individuals themselves are reliable sources to report cognitive processes produced by psychological and physiological anxiety. Although self-report scales have been developed to measure FLCA, the scales are specific to academic settings (e.g., MacIntrye & Gardner, 1989). Items from these scales reflect context-specific anxieties that are not appropriate for non-student populations (e.g., “I tremble when I know that I’m going to be called on in language class”, “I am afraid that my language teacher is ready to correct every mistake I make”; Horwitz, Horwitz, & Cope, 1986, p. 129). McCroskey, Beatty, Kearney, and Plax’s (1985) general communication anxiety measure, the PRCA-24, has been used to measure apprehension in speaking a different language by adding the instructions “WHEN I SPEAK SPANISH” or “WHEN I SPEAK ENGLISH” to the top of the instrument (see McCroskey, Fayer et al., 1985). However, these measures are not specific to a medical context. MacIntyre and Gardner (1991) note “a situational perspective provides the best research approach [to studying foreign language anxiety]” (p. 112). FLCA may manifest itself differently in a medical setting than in other contexts, such as in social interactions.
Measurement tools (e.g., a survey) should be psychometrically sound in order to properly test and explicate a construct (Crano & Brewer, 2002; DeVellis, 2003). A psychometrically sound measure is both reliable (“the proportion of variance attributable to the true score of the latent variable”) and valid (“whether the variable is the underlying cause of item covariation”; DeVellis, 2003, p. 27 and 49, respectively). The proposed scale, the Foreign Language Anxiety in a Medical Office Scale (FLAMOS), is tested for internal consistency, as well as convergent, divergent, and predictive validity. The proposed scale is hypothesized to measure a uni-dimensional construct that is related to, but unique from, measures of general communication apprehension (PRCA-24, McCroskey, Beatty et al., 1985; Communication Anxiety Index, Booth-Butterfield & Gould, 1986; Receiver Apprehension Test, Wheeless, 1975). The measure should also account for participant reported acculturation related to language use (e.g., English vs. Spanish) to establish the predictive validity of the scale. As FLAMOS should address communication behaviors specific only to a medical setting, the scale is not expected to predict acculturation outside of a medical context (e.g., media use or social relations acculturation).
Methods
Participants
Low-income, Spanish-speaking adults (N = 100) were recruited for this study. Participants were predominately female (n = 83) and ranged in age from 18 to 71 years of age (M = 34.85, SD = 11.48). Participants mainly reported Mexico as their country of origin (n = 89) and that they were non-native English speakers (n = 91). Forty-two participants did not speak English, ten spoke “some” or “a little” English, and 25 reported that they have spoken English for three years or more (M = 13.28, SD = 9.75). Nineteen participants reported being U.S. citizens, seven naturalized citizens, six legal immigrants, 25 permanent residents, 25 illegal immigrants, and 18 did not report their citizenship status. The education level of participants ranged from no formal education (n = 2), between grades 1 to 6 (n = 35), grades 7 and 8 (n = 10), grades 9 to 11 (n = 20), and a high school graduate or higher (n = 30).
Procedure
Participants were recruited by two bilingual employees of a university extension program from Lake County, Indiana and were compensated $25 in cash for their participation.Participants completed a consent form, presented in Spanish, and were given a nine-page paper survey in Spanish. Participants were given the option of having the consent form and the survey read to them, as an aid to low-literacy participants. The protocol for this study was reviewed and approved by a university Institutional Review Board.
Scale Creation
The Foreign Language Classroom Anxiety Scale (FLCAS) was created and validated by Horwitz and colleagues (1986) and has been used extensively in FLCA classroom research. Items were adopted from that measure and modified to fit a medical setting rather than a classroom. Longer scales often have higher internal consistency and reliability, but shorter scales are less taxing on participants (DeVellis, 2003). The current scale weighed these trade-offs during construction when considering its intended use. FLAMOS is intended for use in Latino populations who face language barriers and healthcare difficulties, populations that have low health literacy rate (Hoffman-Goetz et al., 2009). As these participants might be overwhelmed or discouraged by lengthy questionnaires, FLAMOS modified 12 of the original 33 items for the current study.