Spanish Content on Hospital Websites: An Analysis of U.S. Hospitals' in Concentrated Latino Communities

Authors


Abstract

This study explores the nature and extent of Spanish language translation of hospital websites. A quantitative content analysis of 121 U.S. hospital websites located in Hispanic communities was conducted. Results indicate that a significant number of U.S. hospitals fail on their websites to offer equal content for Spanish language users. The study also examines several factors potentially associated with the incidence of website translation. First, organizational factors such as hospital size, ownership type and formal statements of diversity commitment are investigated. Second, demographic factors such as county population size and Hispanic population size are also considered. Hispanics' use of the Internet, the digital divide, culture's impact on Internet use, and eHealth provide a background for discussion of the findings.

The importance of eHealth for improving health communication among patients and providers is significant. Currently, health care organizations such as hospitals and public health agencies increasingly use the Internet as a main source of communication and information dissemination (Gibbons, 2005; and Healthy People 2010, 2002). Additionally, eight out of ten Internet users have searched online for health care information in many topical areas including hospitals, according to a survey by the Pew Internet and American Life Project (May 17, 2005). As patients (consumers) use web resources for information, research, and communication, hospital websites are becoming an industry standard (Randeree & Roa, 2004). Yet, hospital websites need to be more useful and relevant to eHealth (Author, 2007).

Ideally, hospital websites should provide information that takes into account the cultural and ethnic diversity of the community in which the physical institution resides. This mandate should be especially true of hospitals in geographic areas that have higher populations of minorities. However, little is known as to the extent to which hospitals are attending to this concern. This issue gains import in the United States, given that one of the largest and fastest-growing minority groups potentially affected are Spanish speaking Latinos. Hispanics (as designated by the U.S. Census Bureau) in the United States account for 46.7 million people or 15.1% of the total population. As the U.S. Census Bureau reported in August 2008, this population is expected to triple with nearly one in three U.S. residents being Hispanic by 2050.

Designing websites to meet the needs of cultural and ethnic diverse populations is an important eHealth objective. As more attention turns to the issue of health care disparities, and culturally competent care, health care organizations and agencies are compelled to consider linguistic and cultural barriers that impede health care access. Governmental initiatives such as the National Standards for Culturally and Linguistically Appropriate Services in Health Care (Office of Minority Health, 2001) standards have considered the role of cultural and linguistic communication barriers that manifest during patient-provider transactions. We contend that cultural and linguistic barriers should also be addressed in the online relationship between hospitals and the minority community that they serve.

In particular, this study focuses on the linguistic and cultural needs of Latino patient populations. As the number of Latino Internet users continues to grow, attention to Internet access disparities should be considered when evaluating the quality of hospital website communication. In other words, after people have the technological access to online communication, the websites' content, though physically accessible, might not be understandable or useful. The current study investigates this premise by analyzing how 121 hospitals in areas with high Hispanic populations use their institutional websites to meet the Spanish language needs of Hispanics and promote cultural diversity.

Literature Review

To better understand how information presented on hospital websites can serve Spanish speaking communities, research on how ethnicity and culture shape Internet use is reviewed. The digital divide, an established research direction on how minorities access and use the Internet, is examined with particular attention to Latino populations' uses of the Internet. Cultural and Internet use is also described in how it impacts eHealth. In this paper, Hispanic and Latino are used interchangeably based on the definition provide by the U.S. Census Bureau (2007). Literature is reviewed on how culture influences Internet use and on the digital divide focusing on Latino Internet use and eHealth.

Digital Divide

The digital divide is generally defined as “gaps in access to computers and the Internet based on ethnicity, education, and household income,” (Rains, 2008, p. 283). Resolving problems central to the digital divide is both a public policy issue (Dupagen & Salwen, 2003; and Newholm, et al., 2008; Prieger & Hu, 2008; Warren, 2007; Wyatt et al., 2005; and Zarcadoolas et al., 2002) and a concern for eHealth (Lorence, Park, & Fox, 2006; Rains, 2008; Wyatt et al., 2005; and Zarcadoolas et al., 2002). Targeting and overcoming digital divide barriers in access to health information for certain at-risk populations can improve information access and healthcare support (Gustafson et al., 2008).

While much of past digital divide research has focused on online access, “postaccess disparities” need more study (Jung, 2008, p. 323; and Morey, 2007). The digital divide is generally categorized as an inequality in having the access to the technology that allows users to physically connect to the Internet while postaccess disparities occur when people have access to the Internet, but may not have adequate language skills, usage skills or content available that relates to their cultural or language needs (Morey, 2007). Understanding the nature of post access issues is imperative to aid consumers of online health information, especially those with medical problems who often search online (Goldner, 2006). To accomplish goals such as gaining health information, the promise of online health information could drive individuals to intense online information-seeking behavior (Jung, 2008, p. 335). However, eHealth can only be leveraged to promote healthcare if a population has the eHealth literacy skills to “read text, use information technology, and appraise the content of these tools to make health decisions” (Norman & Skinner, 2006). Providing easier access to and understanding of online health information through language translation for diverse populations may reduce user frustrations and improve the usefulness of eHealth to help people make healthcare decisions.

Hispanics Online

The percentage of Latinos accessing the Internet is growing. For example, in October 1997, 38% of Hispanics were online compared to 50% of the general population of the United States. Nearly 10 years later, Hispanics online in the U.S. increased by 18% which is comparative to increases by non-Hispanic whites at 21% (Fox & Livingston, 2007). By 2006, at least 56% of Hispanics were online compared to 71% of non-Hispanic whites (Fox & Livingston, 2007). Although Hispanics still lag behind non-Hispanic whites in the traditional formats of Internet use, a variety of studies show Hispanics gaining in technology use including mobile phone usage, text messaging, social media usage, and broadband access. For example, a recent study reported that Hispanics led the way for mobile Internet use and text messaging when compared to whites and blacks (Horrigan, 2008). A Pew Internet & American Life report by Lenhart (2009) revealed that 48% of Hispanic adults are on a social network compared to 31% of white adults and 43% of African-American adults. A 2008 survey by Horowitz Associates claims that 48% of all Hispanics in the U.S. have broadband online connections compare to 57% of the population as a whole (Moss, 2008).

Further, a large segment of Latinos prefer websites in Spanish. For example, data indicates that 52% of U.S. Hispanics who use the Internet prefer to speak English, but 49% either prefer Spanish or use Spanish and English equally (ComScore Media Metrix, 2003). In general, language barriers for Spanish speaking Latinos can be significant. More than 27 percent of Hispanic adults or 7,273,000 individuals have limited English proficiency (LEP) (Brach and Chevarley, 2008). At least one in five residents of the U.S. states of Arizona, California, New Mexico and Texas spoke Spanish at home in 2007. Nationwide, an estimated 35 million use their native language at home (Census Bureau, 2008). Between 2005 and 2025, the Spanish-speaking Hispanic population is expected to grow by 45%, from 27.8 million in 2005 to 40.2 million in 2025 (Roslow Research Group, 2005). Thus, it may be reasonable to conclude that the need for linguistic and culturally appropriate online health information will also increase.

After finding that Hispanics-Americans have an overall, slower Internet adoption rate than Anglo-Americans, Hacker and Steiner (2002) directly point out that as Latino Americans find more Spanish-language websites, their adoption of the Internet as a communication technology will likely increase. Some of the disparity in online behavior may be explained by language dominance among Latinos. Among U.S. Latinos online, 78% are dominant English speakers with 76% being bilingual speakers of English and Spanish compared to only 32% of Spanish-dominant Latino adults online (Fox & Livingston, 2007). It follows that some Latinos may not be going online because of lack of Spanish content: “Nearly three in four offline Hispanics who speak at least some Spanish (71%), say online Spanish content is important. More than half of all offline Hispanics (56%) cite lack of Spanish content as a reason for not going online at home,” (Roper, 2004).

Consumer Informatics and eHealth

An important branch of eHealth research is consumer informatics which evaluates the web-based tools and content consumers use to make decisions about their health. According to Eysenbach (2000), consumer informatics is a branch of medical informatics that “analyzes consumers' needs for information; studies and implements methods of making information accessible to consumers; and models and integrates consumers' preferences into medical information systems.” With 40.2% of Hispanics having visited online content in a health category (ComScore, 2008), Livingston, Minushkin, & Cohn (2008) reported that 35% of Hispanics get consumer health information from the Internet. Yet, there is a dearth of quality online health and medical content for Latinos. Cardelle & Rodriguez (2005) documented that the quality of Spanish health information websites are an emerging disparity in online content. Singh, et.al. (2007) also concur that the vast majority of health Internet web sites worldwide are offered only in English. Many Hispanics are in danger of being left behind not because they reside on the wrong side of the “technology divide,” but because they reside on the wrong side of the “quality gap” (Cardelle & Rodriguez, 2005, p.85). As noted by Berland et al. (2001), “Accessible and high-quality health information on the Internet is important for English speakers; it could be even more useful for Spanish speakers, who face greater barriers to traditional sources of medical care and information.”

In addition to consumer health websites, another source of health information that is gaining in importance is that provided by health care organizations. Fueled by consumer desire to take responsibility for health decisions and pressures experienced by health care organizations to reduce costs, the “information age health care system” has emerged to empower consumers and improve quality of care (Eysenbach & Diepgen, 2001). As noted by Kirshenbaum (2002), health care providers often consider eHealth as a way to efficiently facilitate communication among patients while reducing administrative costs. Further, many consumers look to their health care organizations as trusted sources of local health information that may be web enabled (Biohealthmatics, 2006). It has been argued that Web access to local information in multiple languages including health information is strongly desired by underserved Americans (The Children's Partnership, 2002; U.S. Department of Health and Human Services, 2007). Yet, little research has evaluated health care providers as sources of health information used by consumers (Bodie & Dutta, 2008). Hospitals should be an important focus of consumer informatics and eHealth. For example, Sciamanna, Clark, Houston, & Diaz (2002) identified an opportunity gap or the difference between what people are doing and what people are interested in doing on the Internet. Their research contends that consumers are interested in information provided by hospitals that allow them to evaluate their tests and treatments, schedule an appointment, and to determine the quality of their care.

Whether for-profit or nonprofit, any entity that wishes to communicate with a widened audience needs to have a web presence (Coleman, et al. 2008). In using the Internet to communicate with stakeholders, organizations must plan for complexities of web-based communication. One such difficulty is effective communication geared toward multiple ethnic and racial groups. For healthcare organizations, a concern is that all publics have fundamental healthcare communicated to them. Overall, Latinos concur that online health information helps the comprehension of medical conditions as well as treatments, adds to patient confidence when communicating with doctors, and provides users find medical treatment that would not normally be available (Peña-Purcell, 2008).

However, there are several factors that may inhibit those who seek or need health information provided by health care providers. In addition to access barriers posed by general eHealth literacy issues (Norman, Chirrey, & Skinner, 2002), linguistic barriers may also impede eHealth information. To overcome linguistic barriers to health information among Latinos in the United States, federal agencies such as the Centers for Disease Control and the National Institute of Health have taken the lead and translated their web sites into Spanish. For instance, the National Cancer Institute has a Spanish language website at http://www.cancer.gov/ESPANOL. Yet, while healthcare organization leaders are responsible for stewardship in its communication (Healthcare at the Crossroads Report, 2007), addressing health literacy issues is considered low among health care executives and other managers. Managing the health care information needs of ethnic minorities with limited English proficiency poses many challenges.

Culture & Internet Use

In addition to meeting the linguistic needs of ethnic minorities, a website's content is also meaningful if it takes into account a person's cultural perspective. While navigation behavior within websites are similar across differing cultures; yet, website “content must be in appropriate languages and the quality of the translations must be high” (Zarcadoolas, et al., 2002, p.319). Furthermore, in encouraging repeat visits, a website's usability in navigation and appearance mattered less than its story or written content (Coleman, et al. 2008, p. 194). When applied to healthcare, culturally competent communication is not simply an attribute of the patient-provider encounter (Taylor & Lurie, 2004). It should also be reflected in all aspects of institutional communication. (Taylor & Lurie, 2004). Thus, health care websites provide more utility if content is expressed with the cultural needs of ethnic minority groups in mind.

Latinos seek out culturally sensitive healthcare information. In fact, the lack of ethnically targeted online health information in Spanish may partially explain why first-generation Hispanic immigrants, having lower income and education, turn to other more ethnically sensitive health communication sources such as television and interpersonal communication (Cheong, 2007). While education is the greatest predictor of online health information searches, younger insured second-generation male Hispanics immigrants are more likely of all Hispanics to use the Internet for health information (Cheong, 2007).

In fact, ethnicity and race has been found to influence individuals' preference for community Internet portals. Hargittai (2008) reported that Hispanic students used MySpace rather than Facebook, which are used more by White and Asian students. Website content reproduces existing cultural norms and values of its creators (Brock, 2006), in social networking sites content is produced by users. This may explain why Hispanic usage of online social networking sites at 48% is higher than whites at 31% and African-Americans 43% (see, Lenhart, 2009) revealed that 48% of Hispanic adults are on a social network compared to 31% of white adults and 43% of African-Americans. Whereas social networking site content is developed by its uses (bottom-up), Brock (2006) found that website content development (top-down) reproduces dominate societal patterns that lessen content reflecting minority interests. For instance, Brock (2006) found that website portals such as Yahoo are primarily designed for a White American audience and do not contain African American cultural content like Africana.com which is a website portal designed for African Americans.

As an exploratory investigation, this study seeks to document the nature and extent of hospital Spanish language translation in Hispanic dense communities and determine what factors, if any may influence the incidence of translation. Several factors are considered including organizational properties such as hospital size, ownership type, and evidence of cultural sensitivity as referenced by the inclusion of a formal diversity statement on the website. This study also considers hospital community demographic factors such as the size of the Hispanic population served and county size. The following research questions as offered:

R1: Does hospital size influence the presence of Spanish language translation of hospital websites?

R2: Does hospital ownership type (public, private, not for profit) influence Spanish language translation of hospital websites?

R3: Does the size of the county population influence Spanish language translation of hospital websites?

R4: Does the size of the Hispanic population influence Spanish language translation of hospital websites?

R5: Are hospitals with cultural diversity statements on their websites more likely to translate websites into Spanish?

Methodology

Units of Analysis

In order to develop a sampling frame, hospital websites were selected by the degree of concentration of Hispanic population within a given community. Counties in the United States with over 55% Latino population were selected based on Census Bureau data from 2007. The top 50 counties by largest percentage of Latino population were included in this study. This ranges from 97% Hispanic for Starr County, Texas, to 55% Hispanic in Taos County, New Mexico. Only a few counties in this original list were excluded, because the county was too small to have a hospital or because the nearest hospital in the nearest county was already counted.

Other counties selected for the study were based on large numbers of Latinos from the top 60 counties by population according to 2007 Census data. This ranges from Los Angles Country California with over 4.6 million Latinos to Fairfield County Connecticut with over 133,000 Latinos. For these counties, zip code data was examined to identify areas in the country with a high concentration of Latinos. Hospitals were then selected in those zip codes with a high Latinos concentration or the nearest zip code with a general hospital. Some counties were excluded if the Latino population within the county was too dispersed (less than 25%) when zip codes were examined.

Finally, to insure that variability among states and counties were represented, some counties in the top 100 of Latino population were included if a high concentration (over 40%) could be identified as living in several zip codes within the county. Hospitals in these areas were then included such as hospitals in North Philadelphia for Philadelphia county or Lawrence Massachusetts in Middlesex county.

Hospitals in 62 counties were examined in this study. The state with the greatest number of counties with substantial Latino populations used in this study was Texas (26), followed by California (10), New Mexico (7), Arizona (4), New York and Colorado (3), Illinois and New Jersey (2), and the other states with 1 county each. In order to ensure up-to-date website addresses for the hospitals and organizations, the website links used in study and reported here were re-examined in early January 2009.

A total of 121 hospital websites were identified from hospitals designated by the American Hospital Association directory as “general medical and surgical” facilities. This sample of hospital websites was further analyzed based on several factors including ownership model; number of beds; and urban versus rural demography. Hospital ownership was classified as for-profit, nonprofit and public. This sample, across 13 total states, yielded 58 nonprofit hospitals (47.93%), 36 for-profit hospitals (29.75%) and 27 (22.31)public hospitals. Hospital size was determined by number of beds. Small hospitals (1–100 beds), medium hospitals (101–300 beds), large hospitals (301–499), and extra large hospitals (500 plus beds) were identified (Groeger, Strosberg & Halpern, 1992); Halpern, et.al, 2006). Out of 121 hospitals, 18 were classified as extra large (14.87%), 34 were classified as large (28.09%), 42 were classified as medium (34.71%), and 27 were classified as small (22.31).

Urban and rural hospitals are categorized by the Center for Disease Control's (CDC) standards. The CDC's National Center for Health Statistics has a six-category ranking of urban (i.e., metropolitan) and rural population (as based on Ingram & Franco's (2006) six classification rules which assign counties to the 2006 NCHS Urban-Rural Classification urbanization levels). Large central metro areas are counties with 1 million or more population with the largest principal city or principal city containing at least 250,000 residents. Large fringe metro areas contain nonprincipal cities in counties having 1 million or more residents. Medium fringe metro areas have counties with 250,000 to 999,999 residents. Small metro areas have counties with 50,000 to 249,999 residents. Micropolitan areas have counties with urban clusters of 10,000 or more residents. Noncore areas are defined as all counties not fitting a micropolitan or metropolitan definition.

Nearly half of the hospitals studied or 60 (49.58%) were in large central metros. Cities like Los Angeles, Chicago, New York, San Diego, and Miami had large numbers of Latinos. “Hispanics are highly concentrated; half live in California and Texas, with another 20 percent residing in five additional states with 1 million or more Hispanics: New York, Florida, Illinois, Arizona, and New Jersey. … Almost half reside in a central city, compared to 21 percent of whites” (Pew Hispanic, 2002). Other urban categories held 34 hospitals. There were eight large fringe metro hospitals (6.61%), 22 medium metro hospitals (18.18%), and four small metro hospitals (3.30%) such as El Paso County, Texas; Imperial County, California; or Yuma County, Arizona. Some 27 hospitals were considered rural since there were 16 hospitals in micropolitan (13.22%) and 11 hospitals in noncore (9.09%) counties. Examples of hospitals in rural counties were Starr County, Texas; Brooks County, Texas; or Mora County, New Mexico.

Website Spanish Translation Classification

Spanish classification rules were inductively derived by visiting a variety of hospital websites resulting in seven classification types. The qualitative analysis of Spanish content for any website identified as being mirrored, partial or having an online translation application link were evaluated by one author who is a native Spanish speaker, and another author who is also fluent in Spanish. A description of the data classification rules for Spanish language translation is presented in Table 1.

Table 1.  Spanish Language Classification Rules
No. & NameClassification Rules
1. MirroredThe hospital's website is mirrored in Spanish (or reproduced-not necessarily word for word, but an adequate translation of meaning)
2. PartialPartial site in Spanish – a variety of different pages are in Spanish, but not all the content is mirrored
3. Translation Application LinkLink to online translation application such as Babel Fish
4. Spanish Health Info LinkLinks are provided to other websites with health/medical information in Spanish for example MedlinePlus (http://medlineplus.gov/spanish/)
5. Call Hospital Translation ServicesProvides a hospital number to call for onsite translation services
6. No Spanish on the WebsiteNo Spanish language content on Website
7. Spanish Pages Under ConstructionThe Spanish language web pages were identified as under construction

Data Analyses & Coding

Once the sampling frame was determined, a content analysis was performed. Content analysis is a descriptive methodology suited to identifying “patterns, frequencies, or potential categories” with respect to social phenomena (Carlson, 2008, p. 105) and is specifically useful in analyzing written text in technical communication research (Thayer, et al., 2007). Content analyses were conducted for type Spanish language translation and the presence of cultural diversity statement on hospital websites. The authors examined all home pages and all internal links within the site. Since web content may change, website pages were copied and collected in a 3-day period.

Spanish Language. Each hospital website in the sampling frame was content analyzed for type of translation offered and percentages for each of the seven types of translation was determined. Inter-rater reliability based on two coders yielded a Cohen's Kappa coefficient of .89. Coding reliability was calculated using Kappa as a measure of agreement. Cohen's Kappa values greater than .75 indicate excellent agreement beyond chance. Next, translated classification types were further collapsed into two distinct categories (Spanish language translation or no Spanish language translation) for the purposes of further measurement. Websites classified as mirrored, partial, having an online translation application link and Spanish health info link were classified as having Spanish language translation. The last three categories (call hospital translation services; no Spanish on the website; and Spanish pages under construction) were classified as having no Spanish language translation. Next, Chi square tests were conducted to see if there was a relationship between Spanish language translation and the following units of analysis: hospital size; ownership type (public, private, and non-profit); county size1; and Hispanic population size.

Cultural Diversity Statement. A content analysis based on defined units of analysis was conducted to determine the percentage of hospitals with formal diversity statements. Many nonprofit hospitals have a mission of charity based on religious affiliations and doctrine. In identifying cultural diversity statements, the authors discussed whether or not statements concerning poverty and charity-based care should be defined as diversity. The authors agreed that a mission to serve the poor did not equate for this study to cultural diversity related to ethnicity and culture. After this agreement, coding for cultural diversity statements is exclusively defined as a declaration affirming differences in culture, ethnicity, or race. Interrater reliability based on two coders yielded a Cohen's Kappa coefficient of .92. This Cohen's Kappa value is well above .75 indicating excellent agreement beyond chance.

Results

This study was designed to determine if Spanish language translation of hospital websites was related to characteristics such as hospital size, hospital ownership type, metropolitan population size, Hispanic population size or a hospital's formal statement of cultural diversity. Content analyses of Spanish language translations of and cultural diversity statements on hospital websites are reported in tables 3 and 4 respectively.

Table 3.  Cultural Diversity
No. & NameOwnership Type 
For-ProfitNonprofitPublicTotal
1. Had CD statement1217534 (28.09%)
2. No CD statement24412287 (71.90%)
Total   121
Table 4.  Spanish * Hospital Size Crosstabulation
   Hospital SizeTotal
   Extra Large & Large HospitalsMedium HospitalsSmall Hospitals
SpanishNo TranslationCount16251859
  Expected Count25.420.513.259.0
  % within Hospital Size30.8%59.5%66.7%48.8%
 Spanish TranslationCount3617962
  Expected Count26.621.513.862.0
  % within Hospital Size69.2%40.5%33.3%51.2%
TotalCount524227121
 Expected Count52.042.027.0121.0
 % within Hospital Size100.0%100.0%100.0%100.0%

Slightly more than half or 51% (n = 62) of the 121 hospitals had Spanish language translation on their websites. See table 2. Twelve, or 9.91%, of hospital websites offered mirrored translation. Thirty-seven, or 30.57%, of hospitals offered partial Spanish translations with content ranging from embedded Spanish health encyclopedias, video, and PDFs with particular health and disease information and administrative information such as admissions instructions and hospital directions In a few cases (10.73%), hospitals offered links to third party providers of translation (e.g. Bablefish) or sites that offered translated health information.

Table 2.  Spanish Language
Classification CodeNo.
1. Mirrored12(9.91%)
2. Partial37(30.57%)
3. Translation Application Link3(2.47%)
4. Spanish Health Info Link10(8.26%)
5. Call Translation Services2(1.65%)
6. No Spanish on Website54(44.62%)
7. Spanish under Construction3(2.47%)
Total121

Further, a vast majority of hospital websites 87 (71.90%) did not contain a cultural diversity statement. On the other hand, 34 (28.09%) hospitals did have cultural diversity statements on their websites. These numbers are further broken down by hospital ownership type (for-profit, nonprofit, and public) in table 3.

Of the five research questions posited, only questions one and five were significant. Research question one asked if hospital size influenced the presence of Spanish language translation of websites hospital websites. In this study, hospital size did make a difference in whether or not Spanish language was present on a website. Chi square tests reveal that hospitals of larger size are more likely to have a Spanish language presence on their websites, X2(2, N = 121) = 12.194, p = .002. In this sample, 69.2% (36) extra large and large hospitals, 40.5% (17) of medium hospitals and 13.8% (9) of small hospitals offered some degree of translation. The results are presented in table 4.

Research question five asked if hospitals with cultural diversity statements on their websites were more likely to translate their websites in Spanish. Observed data indicates that there is a difference between hospital website cultural diversity statements and Spanish language translations on hospital websites, X2(1, N = 121) = 4.129, p = .042. Thus, having a cultural diversity statement are more likely to have a Spanish language translation.

The remaining three research questions demonstrated no significant difference. Research question two asked if hospital ownership type (public, private, not for profit) influences Spanish language translation of hospital websites A Chi Square test, X2(2, N = 121) = .143, p = .931, failed to reject the null hypothesis; thus, there is no difference between hospital type and Spanish language translation on hospital websites.

Research question three asked if the size of the regional population influenced Spanish language translation of hospital websites. In comparing large central metropolitan area hospitals with nonlarge central metropolitan hospitals, the Chi Square test , X2(1, N = 121) = 1.403, p = .236, failed to reject the null hypothesis. There was no difference in Spanish language translation among hospitals in large central metropolitan geographic areas and lesser populated geographic areas.

Research question number four asked if the size of the Hispanic population influenced the Spanish language translation of hospital websites. Again, no difference was found. Chi Square test, X2(2, N = 121) = 0.227, p = .893, failed to reject the null hypothesis. Thus, Hispanic population bore no difference on whether or not hospital websites have Spanish language translation.

Discussion

Research is calling for new approaches and policies to provide additional and appropriate health information for underserved and vulnerable populations to facilitate better-informed healthcare decisions, including finding appropriate medical care (Chang et al., 2004; Kreps, 2005). Hispanics perceive that using online health information can improve their understanding of medical conditions and treatments, leading to increased confidence when talking to doctors about health problems, and to better-quality medical care (Baker, et al., 2003; and Peña-Purcell, 2008). The current study adds to the body of knowledge for how eHealth can be better utilized in reducing health disparities for underserved and vulnerable populations. More specifically, these findings illustrate a gap in U.S. hospitals' ability to provide informed healthcare information through their websites for a growing cultural community, Latinos.

In this study, very few hospitals offered a mirrored website. Moreover, the quality of mirrored websites varied. For example, some websites, such as St. Vincent's Hospital and Bronx Lebanon in New York City, Sharp Health in San Diego, and Park Plaza in Houston offered complete and localized translation of their website from English to Spanish. However, some websites such as Hoboken University Medical Center, simply offered a complete website translation through Google's automatic translation function. Unfortunately, this nature of translation may be confusing for Spanish features in that it cannot account for contextual meaning. Websites offering partial translation made use of different information structures. For example, some partially translated websites embedded video content or offered PDF brochures with particular health and disease information that could be utilized by Spanish speaking users. In Texas, for example, Corpus Christi Medical Center in Corpus Christi and Saint David's Medical Center of Austin both offered an extensive list of diseases with their corresponding explanations in Spanish. Many partial websites also offered administrative information. However, some partially translated websites seemed to have Spanish pages that appear to link to Spanish content. Yet, when these links are accessed, the user is taken to web pages solely in English. Some links were simply not operational. In other instances, health information content was simply basic. For example, some partially translated websites offered less detailed health information in Spanish than what was offered in English. In other cases, partially translated websites only provided information about the hospital location, policies, or services but fail to provide health or medical information that was readily available to English users.

Many hospitals with partial translation also used the same website template or content management system, which is simply a way to easily publish websites for organizations. A content management system for websites provides a ready-made computer application that automates the construction of a website. These content management systems also allow the hospital website to pull information from health information vendors' such as EBSCO Host databases. Health information content from these databases is also available in Spanish. In the hospital websites reviewed, this process is often used for Spanish health information. It is interesting to note that some hospitals using the same website template from EBSCO failed to provide Spanish content, even though EBSCO provides a Spanish content service.

This study also sought to determine which factors may account for the incidence of website translation. Surprisingly, hospital websites in geographic areas with significant Latino populations across the U.S. lacked relevant, translated information. However, large size hospitals were found to be exceedingly likely to have Spanish language translation, as defined by this study, on their websites. Additionally, hospitals presenting a diversity statement on their websites more often than not had Spanish language translations.

There are two potential explanations of the results. First, large and extra large hospitals may have the resources necessary to provide Spanish translation on their websites. Unlike larger hospitals, small to medium-sized hospitals in this sample were less likely to invest in translation. Yet, providing health information on hospital websites for speakers of Spanish can be done without great cost. There are excellent medical and health information websites in Spanish provided by government and nonprofit associations (see table 6 which provides 21 links referenced by various hospitals in the current study). These 21 listed websites provide free health information in Spanish. Surprisingly, only 10 or 8.26% of 121 website reviewed had links to these free high-quality health information websites in Spanish. Our results are consistent with observations made by Fulda, et.al (2004). In their study of hospitals in the South Central Region of the U.S., the Medical Library Association's most useful health websites were not widely used on hospital websites. They concluded that there may be a lack of awareness of this highly selective but reputable list, especially among hospital Website coordinators.

Table 6.  Links to Externally used Websites with Health Information in Spanish
Alzheimer's Association - http://www.alz.org/espanol_recursos_para_los_latinos.asp
American Academy of Allergy, Asthma, & Immunology - http://www.aaaai.org/espanol/
American Cancer Society - http://www.cancer.org/docroot/ESP/ESP_0.asp
American Diabetes Association - http://www.diabetes.org/espanol/
American Heart Association - http://es.aha.drtango.com
American Lung Association - http://www.lungusa.org/site/pp.asp?c=dvLUK9O0E&b=33214
Arthritis Foundation - http://www.arthritis.org/espanol/
Center for Disease Control and Prevention - http://www.cdc.gov/spanish/
Epilepsy Foundation - http://www.fundacionparalaepilepsia.org/
Healthy Day - http://www.healthday.com/espanol/
March of Dimes - http://www.nacersano.org
Medline Plus - http://medlineplus.gov/spanish/
Muscular Dystrophy Association - http://www.mdaenespanol.org/
National Cancer Institute - http://www.cancer.gov/espanol/
National Center for Health Statistics - http://www.cdc.gov/nchs/about/nchs_en_espanol.htm
National Institutes of Health - http://salud.nih.gov
National Multiple Sclerosis Society - http://www.nationalmssociety.org/informacion-en-espanol/index.aspx
National Stroke Association - http://www.stroke.org/site/PageServer?pagename=espanol_que_es
U.S. Consumer Product Safety Commission - http://www.cpsc.gov/cpscpub/spanish/spanish.html
U.S. Department of Health and Human Services - http://www.healthfinder.gov/espanol/
U.S. Food and Drug Administration - http://www.fda.gov/oc/spanish/

Financial subsidies help healthcare organizations develop and advance community eHealth initiatives (Mostashari, Tripathi, & Kendall, 2009). Still, a hospital's organizational culture and administrative direction can set a tone that supports or hinders progress on eHealth for multicultural constituencies. Some hospitals' management simply may not value eliminating eHealth disparities because it is not perceived as generating return on investment. Aiding low-income communities with less insurance can be a financial burden on hospitals. This scenario is even seen in not-for-profit hospitals with considerable charitable and government subsidies. According to Jacoby & Warren (2006) these not-for-profit hospitals may show “aggressive businesslike behavior” that is inconsistent with the public image of how charitable organizations should act in providing access to healthcare for at-risk communities.

A possible solution in developing better eHealth for at-risk communities is to build alliances. Along with organizational leadership and capital, another important resources to support and advance eHealth are community and localized county government, which together shore up coalitions to move forward with eHealth initiatives (Frisse, 2005). A push to eliminate eHealth disparities may need to take advantage of this model that combines organizational leadership with community and government support.

A second explanation for the results may be attributed to organizational leadership that values cultural sensitivity resulting in proactive initiatives to eliminating eHealth disparities. With this explanation, one might argue that organizations that fail to provide adequate website translation simply lack awareness of the importance of providing culture- and language-specific content for minority populations. Not surprising, the largest percentage of institutions that did not offer a translation (48.1%) or a diversity statement (81.0%) were public hospitals. Public hospitals are dependent on public sources of financing that are sensitive to economic downturns and often serve uninsured and low-income patients. Their economic status may make it difficult to afford website localization. However, what is surprising is that in this sample, there is no mention of commitments to cultural diversity given the communities that they serve. Further, failure to provide information regarding language translation services (in Spanish or English), on a public hospital website, may violate the spirit of federal mandates that promote cultural and linguistic competence in health care.

The use of online automatic translation application for websites presented many problems with the translation integrity. While this may be an inexpensive and easy way to present Spanish translations on a hospital website, translation inaccuracies in health information have the potential to cause harm, especially if people utilize health information incorrectly due to poor translations. Any health or medical information presented to users should be well developed and accurate. This problem position lends support for Morey's (2007) call for “cultural usability” in which information designers should build eHealth artifacts that take into account online health information needs of ethnic communities that hospitals serve. The Latino language community is highly diverse. Cultural usability would ensure that translations that are culturally sensitive, not just word-for-word translations. Further, cultural usability initiatives would respond to the particular health and cultural needs of a community. For example, a website giving information of better nutrition would refer to the foods actually used by a particular community.

Strength & Weaknesses

As an exploratory study with public policy implications, this is one of the first studies of how a major segment in the U.S. healthcare system, hospitals, uses online presence to serve Hispanic communities. While the current state of Spanish language representation is found to be low on hospital websites that should be serving Hispanic populations, the analysis provides insights into improving this public health information gap. Using hospital websites as an artifact of study provides a public point of access; yet, a weakness is that the information analyzed had no direct information from any of the hospitals' internal stakeholders. As such, these hospitals' representatives could not provide reasons for the lack of Spanish language or cultural diversity statements.

Future Research Directions

Hospital websites should be studied to make them more effective in reaching the local community, providing easier access to culturally sensitive information, and to promote specific health information to targeted populations. A specific follow up to this study can be to directly contact the 121 hospitals to ascertain reason for having and not having Spanish language presence on their institutional websites. More in-depth descriptive research could help theory develop and result in more generalizable findings. For example, interviews with Hispanic speakers can help to determine what content is relevant or more important for translation on a hospital website. This type of dataset could also help set a basis for understanding if hospitals with specific language translation for targeted ethnic communities lead to more trust of the institutions among ethnic community members. Trust is strong predictor of a person's likelihood of using online technology (Gefen, Karahanna, & Straub, 2003). Similarly, interviews with hospital administrators may help to establish reasons for or against web site language translations. Here insight into how leadership teams influence or inhibit advances in eliminating eHealth disparities can be better understood.

Future studies can examine the effectiveness of using social media, text messaging, other features that can link hospitals to their local communities. Further research investigating how existing online health information is presented for Hispanic populations is vital. In particular, the use of rich media that offers audio and visual information in Spanish is highly encouraged. Future research and initiatives, especially in reaching audiences for minority healthcare, need to take advantage of established community entities such as hospitals and their websites as communication vehicles to promote health information. Even government agencies need to be more cognizant of this. The U.S. Department of Health and Human Services' Office of Minority Health (OMH) reflects a commitment to the growing need for devoting more resources to bridge gaps in health care to minority communities. While the OMH mission is to “improve and protect the health of racial and ethnic minority populations through the development of health policies and programs that will eliminate health disparities” (Office of Minority Health, 2009), there is a noticeable absence of web-based initiatives and strategies for reducing disparities in healthcare; even though, ironically OMH has a Spanish version website. The influence of public policy and legislation in eHealth should be studies in relationship to website content for eliminating health disparities.

The lack of focus on communication technology strategies for culturally diverse communities can send a cynical message about commitments to diversity. Future research can address lack of commitment by investigating the reasons health organizations do not have more specific user-friendly online content for multicultural populations. While the current study focuses on identifying a particular language differential reflective of the eHealth literacy needs of the Hispanic population, opportunities exist for widening the scope of focus to other cultural groups based on abilities, age, sexual orientation, and other targeted populations. Another reason for the lack of more culturally targeted content could be resources. Based on having a more robust network of employee skills and financial resources, hospitals owned by healthcare conglomerates and healthcare networks may have more website translations for minority population than smaller institutions. These types of resource distinction should be considered in future research.

Conclusion

The study showed that U.S. hospitals fail on their websites to offer equal content for Spanish language users. These results are consistent with past research in eHealth disparities. Racial and ethnic healthcare disparities are recognized as an urgent national problem that has risen to the point beyond documenting the phenomena to establishing ways to reduce the inequalities (Gibbons, 2008, p. 7–8). Larger hospitals' websites, mostly likely due to increased resources, offered Spanish translations. Furthermore, hospital websites with cultural diversity statements, indicating an institutional multicultural awareness, often have Spanish language translations. This indicates that increased resources and cultural sensitivity awareness might be central to progress in diminishing eHealth disparities.

Notes

  • 1

    To facilitate appropriate cell counts six categories were collapsed into two categories of urban and rural distinction: large central metros which accounted for 60 (49.58%) hospitals; and, non-large central metros which accounted for 61 (50.41%) hospitals.

Ancillary