- Email: [email protected]
Spanish language content on reproductive endocrinology and infertility practice websites
Spanish language content on reproductive endocrinology and infertility practice websites Laura C. Londra, M.D.,a Kyle J. Tobler, M.D.,a Kenan R. Omurtag, M.D.,b and Michael B. Donohue, M.B.A.c a
Division of Reproductive Endocrinology and Infertility, Department of Gynecology and Obstetrics, Johns Hopkins School of Medicine, Lutherville, Maryland; b Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, Washington University School of Medicine, St. Louis, Missouri; and c Freelance Enterprise Resource Planning (ERP) Software Consultant, Association of Reproductive Managers, Baltimore, Maryland
Objective: To analyze the use of Spanish language translation on the websites of reproductive endocrinology and infertility (REI) practices in the context of evidence of underuse of infertility services by minority populations. Design: Cross-sectional survey of websites from REI practices. Setting: Not applicable. Patient(s): None. Intervention(s): None. Main Outcome Measure(s): Assessment of the relationship between having a Spanish-translated website and REI practice characteristics. Variables included concurrent use of social media, size of the practice, Spanish-speaking practitioner in the practice, being a private or a university-based practice, being in a mandated insurance state, and being in an area with different levels of percentage of Hispanic population, adjusted for annual income levels of the population. Result(s): Of the 376 REI practice websites analyzed, 101 (27%) offered at least some information in Spanish. We identified 97 Spanishspeaking practitioners at 71 REI practices. Having a Spanish-translated website was significantly associated with the practice's use of social media, having an international/out-of-town web page, and having a Spanish-speaking physician in the practice. The size of the practice, as measured in number of cycles reported per year, was not associated with having a translated website. In practices located in the top 60 metropolitan areas by Hispanic population, the odds of having a Spanish-translated website were only related to the percentage of Hispanic population after adjusting for state-mandated insurance and average annual income level of the Hispanic population. Sixty-six of the websites with Spanish-translated content had been automatically translated. An additional eight websites were partially translated automatically. Conclusion(s): REI practices in metropolitan areas with a higher percentage of Hispanics were more likely to reach out to this minority population by translating their website content into Spanish. These practices were also more likely to use social media. Future studies are needed to determine whether the availability of Spanish language content on REI websites is associated with increased use of reproductive services by this minority population. (Fertil SterilÒ Use your smartphone 2014;102:1371–6. Ó2014 by American Society for Reproductive Medicine.) to scan this QR code Key Words: Minority health, practice management, health communication, consumer health and connect to the information, Hispanic Americans, translating Discuss: You can discuss this article with its authors and with other ASRM members at http:// fertstertforum.com/londral-spanish-language-rei-practice-websites/
T
he U.S. Hispanic population surged by 43% between 2000 and 2010, rising to 50.5 million in 2010 (1). The U.S. Census Bureau projects that the economic power of
Hispanics in the United States will rise to more than $1.2 trillion in 2012 (2), larger than the entire economies of all but 13 countries in the world (3). This demographic growth is paralleled by
Received April 9, 2014; revised and accepted July 25, 2014; published online September 11, 2014. L.C.L. has nothing to disclose. K.J.T. has nothing to disclose. K.R.O. has nothing to disclose. M.B.D. has nothing to disclose. Reprint requests: Laura C. Londra, M.D., Division of Reproductive Endocrinology and Infertility, Department of Gynecology and Obstetrics, Johns Hopkins School of Medicine, 10751 Falls Road, Suite 280, Lutherville, Maryland 21093 (E-mail: [email protected]). Fertility and Sterility® Vol. 102, No. 5, November 2014 0015-0282/$36.00 Copyright ©2014 American Society for Reproductive Medicine, Published by Elsevier Inc. http://dx.doi.org/10.1016/j.fertnstert.2014.07.1243 VOL. 102 NO. 5 / NOVEMBER 2014
discussion forum for this article now.*
* Download a free QR code scanner by searching for “QR scanner” in your smartphone’s app store or app marketplace.
the increasing rate at which the Hispanic population is accessing medical information online (4). Yet studies have shown that utilization of assisted reproductive technology (ART) services by Hispanic women has not increased, even when economic barriers are minimized (5). In a study of military personnel, despite enhanced access to reproductive services, Hispanic patients did not use ART to the same extent as other ethnic groups (6). Other factors, such as social and cultural influences, 1371
ORIGINAL ARTICLE: ENVIRONMENT AND EPIDEMIOLOGY may explain the significantly lower use of reproductive services compared with the white and African American populations (5). After accounting for cost of treatment, Hispanics indicated that they were less inclined to seek out infertility treatments because of language or cultural barriers, lack of continuity with providers, difficulty scheduling visits and procedures, and a perceived decreased availability of services (7). There is a need to identify and remove barriers faced by Hispanic couples who wish to access information about reproductive services and ART. We hypothesized that communication of reproductive health services to Hispanic populations in the United States is limited by an overall lack in quantity of Spanish language health information delivered via reproductive endocrinology and infertility (REI) practice websites. In addition, we hypothesized that practices that provide Spanish language translations would also have similar features, such as Spanish-speaking physicians. We tested these hypotheses by analyzing the websites of REI practices in the United States for their use of Spanish language translation and its correlation with various practice characteristics and the Hispanic demographics of the practice locations.
MATERIALS AND METHODS Data Collection
The Johns Hopkins University School of Medicine Institutional Board Review (IRB) determined that this type of study is exempt from IRB review. From September 2013 to October 2013, we analyzed the websites of all practices that report to the Society for Assisted Reproductive Technology (SART), which represents approximately 88% of all REI practices in the United States (8). The SART website was used to identify all infertility clinics listed under the ‘‘Find a Clinic’’ tab, which is part of a public domain of the SART website. Using the name and address information from the SART website, Google and Bing searches were used to find the clinic websites. We collected data on practice characteristics based on the information available on the practices' websites, including the availability of Spanish-speaking physicians and their reported level of fluency, use of social media, whether the practice setting was private or university based, existence of a web page directed to international and outof-town patients, additional languages translated, and whether or not the practice was located in a state with mandated insurance coverage for reproductive services. The size of the practice was estimated by using the number of IVF cycles reported per year; these data were then used to perform logistic regression to estimate the odds of having a translated website for each increment in 100 reported cycles per year. For Spanish-translated websites, we distinguished between automatic translation and human translation.
Statistical Analysis Presence of Spanish translation on the websites was used as the dependent variable. The availability of Spanishspeaking staff within the REI practice, the use of social media, 1372
private or academic practice setting, and location in a state with mandated insurance coverage for reproductive services were treated as independent variables in univariate logistic regression analyses, accounting for within-state correlations of characteristics. The size of the practice was evaluated by using the number of reported IVF cycles per year as a surrogate for the number of patients treated and the wealth of the practice. Logistic regression was conducted using increments of 100 cycles per year, clustered on state. We then estimated the odds of having a translated website with each increment in the number of annual cycles per practice. Additionally, the top 60 U.S. metropolitan areas by Hispanic percentage and total population (9) were correlated with the information obtained from the REI practice websites as follows: we noted all of the practices that were located within the 60 largest metro areas by Hispanic population. Then the percentage of Hispanic population from these metropolitan areas and their average annual income levels were categorized separately and clustered by state. Categories of percentage of Hispanic population were as follows: 0–10%, >10%–30%, and >30%. Average annual income categories were (in U.S. dollars) 30,000 to 40,000, >40,000 to 50,000, >50,000 to 60,000, and >60,000 to 70,000. These independent variables were used to conduct logistic regression analysis, accounting for within-state correlation of website characteristics, to identify factors associated with the presence of REI practices that have a translated website, adjusting for category of percentage of Hispanic population, category of annual income, and location in a state with mandated insurance for reproductive services. Analyses were performed using STATA (v 11, Statacorp); P< .05 was considered statistically significant.
RESULTS A total of 376 of 396 SART member practices were included in the analysis. Twenty clinics in the SART registry were excluded either because they did not have a website or because the website was a duplicate with only the practice location and/or URL information being different. Upon review of the 376 websites, 104 (28%) contained content with foreign language translation. The overwhelming majority (97%) of translated websites provided a Spanish language translation. Only three websites did not offer Spanish; those sites were translated into Chinese and Japanese. Among the 101 REI websites that offered Spanish translation, 27 used a human translator and 66 had installed a widget that allows users to automatically translate the website material into the target language. Eight practices used a combination of both, using a launch or splash page with the essential details translated into Spanish. The combined option provided potential patients with a localized introduction to the practice, plus access to the entire site (note: localization is defined as translation and adaptation of content for a particular audience). Nineteen of the 27 practices that employed professional or human translators provided a complete or nearly complete translation of their website content. Results from logistic regression analyses evaluating the association of Spanish language translation with various VOL. 102 NO. 5 / NOVEMBER 2014
Fertility and Sterility®
TABLE 1 Association between practice characteristics and having a Spanish-translated website. Translated website, No translated website, n [ 101 (%) n [ 275, (%)
Practice characteristica In a state with mandated insurance for infertility services Uses social media Includes a Spanish-speaking physician Has a web page directed to international/out-of-state patients University based Private practice a b
63 (62.3) 80 (79.2) 32 (31.7) 41 (40.6) 20 (19.8) 81 (80.2)
138 (50.2) 165 (60) 39 (14) 32 (11.6) 58 (21.1) 212 (77.1)
OR 1.64 2.50 2.80 5.18 0.92 1.20
95% Confidence interval P valueb 0.74–3.62 1.71–3.76 1.72–4.56 2.9–8.97 0.52–1.63 0.68–2.10
.215 < .001 < .001 < .001 .521 .517
Reference category for each predictor is not having the characteristic. Logistic regression adjusting for within-state correlation of website characteristics.
Londra. Spanish language in REI websites. Fertil Steril 2014.
independent variables for all websites and for those in the top 60 metropolitan areas in the United States are shown in Tables 1 and 2, respectively. Logistic regression on number of annual cycles reported per practice using increments of 100 cycles per year, clustered on state, did not show a significant association with having a translated website (OR, 1.05; 95% CI, 0.99–1.11; P¼ .076). Upon examination of the practices located in the 60 largest metropolitan areas in the United States (9), in general, the percentage of the local Hispanic population was associated with practices having Spanish translation of their websites (Fig. 1). Exceptions did exist: 85% of practices in the Boston metropolitan area had websites with Spanish translation, despite Boston having only a 10% Hispanic population. Annual income categories of the Hispanic population were analyzed using logistic regression, and the income category 4 (>60,000–70,000 U.S. dollars) was considered the reference category. There was a statistically significant difference across the three highest income categories (>40,000 U.S. dollars per year) in Hispanic population income categories. The annual income category 3 (>50,000–60,000 U.S. dollars) and annual income category 2 (>40,000–50,000 U.S. dollars) were associated with increased odds of availability of REI websites with Spanish translation in the area (OR, 7.20; 95% CI, 1.89–27.55; P¼ .004; and OR, 4.95; 95% CI, 1.17–20.99, respectively), adjusted by state where the practice is located. As shown in Table 1, 71 physicians indicated on their practice websites that they had some degree of Spanish proficiency. Physicians used a variety of terms on their web
TABLE 2 Likelihood of having a Spanish-translated website by categories of percentage of Hispanic population in the 60 largest metropolitan areas in the United States.a OR
95% Confidence interval
P value
Reference 4.26 7.43
– 1.55–11.7 2.83–19.48
– .005 < .001
Hispanic population (%) 0–10 11–30 >30
a Logistic regression analysis was used to calculate OR, adjusting for categories of average annual income level of the Hispanic population and state-mandated insurance, and accounting for within-state correlation of responses.
Londra. Spanish language in REI websites. Fertil Steril 2014.
VOL. 102 NO. 5 / NOVEMBER 2014
page to describe their proficiency in Spanish. The selfdescribed capability was as follows: ‘‘Spanish’’ or ‘‘Spanish spoken’’ (32%); ‘‘Fluent in Spanish’’ (30%); ‘‘Graduate of a university in a Spanish-speaking country’’ (13%), usually as part of a physician profile; ‘‘proficient/conversant in Spanish’’ (4%); bilingual or multilingual (4%); and other (16%). Having a Spanish-speaking physician on staff was also found to be associated with practices translating their website into Spanish. Our analysis also found that 245 REI practices (65%) engaged in social media through their websites. Of those, 97% of them displayed links to Facebook, followed by Twitter (75%), YouTube (43%), Googleþ (28%), LinkedIn (19%), RSS (19%), and Pinterest (14%). Practices engaging in social media had 2.5-times higher odds of having Spanish-translated content than websites not using social media (Table 1). Seventy-three REI practices made an effort to serve patients outside their local area by offering an ‘‘international or out-of-town’’ web page. Interestingly, almost half of these websites offered no translation to a foreign language (Table 1). Most of these practices included on these pages steps on how to make the IVF experience a successful one, providing new patient forms and listing the necessary medications and steps to take before arriving at the office. A few practices used their websites to promote overseas recruiting trips to conduct seminars for prospective patients.
DISCUSSION We found that just over one-quarter of REI practices sampled in the United States had Spanish-translated content on their practice websites (Fig. 2). Generally, practices located in areas with a large Hispanic population were more likely to have Spanish-translated website content, suggesting that these practices recognize the need for outreach to the growing number of Hispanic health consumers in the United States. The estimated annual income levels of the Hispanic population appeared to have an influence only in the middle categories (annual income >40,000–60,000 dollars per year), likely reflecting the fact that patients with a higher income might be more likely to be more educated and therefore less sensitive to the lack of a translated website when looking for infertility care services. 1373
ORIGINAL ARTICLE: ENVIRONMENT AND EPIDEMIOLOGY
FIGURE 1
Percent of REI practices in U.S. metropolitan areas providing Spanish language translation on their websites, by Hispanic population. The total number of Hispanics in each city is represented by the size of each sphere. Placement of each sphere on the graph illustrates the relationship between the size of the Hispanic population and the percentage of practices that have Spanish-translated websites for each city. For example, 100% of REI practices in the Ft. Lauderdale area (which is 25% Hispanic) translated their websites into Spanish, compared with 85% of REI practices in Boston that translated their websites into Spanish, despite having a population that is only 10% Hispanic. Owing to space constraints, only the 10 cities most and least likely to translate their website are included in the figure (in addition, New York, Los Angeles, and Houston were selected for comparison purposes). Demographic data are from the Pew Research reference (9). Londra. Spanish language in REI websites. Fertil Steril 2014.
Machine translation is cost-effective and does a remarkable job of producing instant translations in an extensive array of languages. However, machine translators often
struggle with nuances of human language (Supplemental Table 1). Machine translation also ignores the fact that U.S. Hispanics come from 22 different countries, with
FIGURE 2
Most frequently offered language translations on U.S. REI practice websites. Each unit represents a link to a individual website translation, except for the Google translate link, which supports 80þ languages, including Spanish. Londra. Spanish language in REI websites. Fertil Steril 2014.
1374
VOL. 102 NO. 5 / NOVEMBER 2014
Fertility and Sterility® various language preferences and degrees of U.S. acculturation. A study on machine translation and gender found that source texts used with Google's translation algorithms lead to unintentional sexism, such as overly assuming that doctors are men and teachers are women (10). The issue of translation in health care was exemplified in 2009, when New York pharmacies were required to provide multilingual medicine labeling; the effort to comply revealed the difficulty in successfully incorporating automatic translation into medical practice. A subsequent study in 2010 showed that machine translations were associated with up to 50% of prescription translations with serious errors (11). Therefore, potential inaccuracies in machine translations may not only misinform patients but also damage the image of the practice. While physicians recognize the need to reach out to Hispanic populations through Spanish translation of their practice websites, the language barrier may continue to pose a challenge to providing care to Hispanic patients. A study by the Pew Hispanic Center found that a mismatch between Hispanic patients' primary language and the language spoken during their doctor appointments lowered the patientreported satisfaction with their care. Up to 30% of Spanish speakers whose appointments were usually conducted in English rated their care as fair to poor, compared with only 19% percent of those whose appointments are conducted in Spanish (12). For those physicians who cannot adequately explore the complexities and nuances of patient interactions in Spanish, an alternative may be to make use of the skills of a trained medical interpreter or translator. Use of remote interpreting services (such as telephone interpreters) also has proven to be very successful (13, 14). We found that two-thirds of REI practice websites engaged in social media, dramatically higher than the less than a third of practices reported just 2 years ago. The use of social media to connect with patients, regardless of ethnicity, appears to be increasing in importance among medical practices. Previous research (15) has shown a positive association between clinic size and the use of social media; we did not find an association between a translated website and clinic size, but we did find a correlation between the use of social media and having a translated website. A recent Pew Report found that ‘‘Hispanics are the most active group on social media networking sites, with 80 percent of Hispanic adults in the U.S. using social media, followed by African Americans (75 percent) and Caucasians (70 percent)’’ (16). It remains unclear whether Hispanics use social media in the context of their health care. Interestingly, practices that had a translated website were also more likely to use social media and to display an out-of-town page for potential patients from other states or other countries. These findings suggest that REI practices in the United States are increasingly open to interact with and serve patients beyond the local residents regardless of clinic size as estimated by number of cycles per year. Overall, more than half of all REI practices neither translate their website nor identify a Spanish-speaking physician on staff. Population demographics may explain why some practices concentrate on communicating with the nonVOL. 102 NO. 5 / NOVEMBER 2014
Hispanic white population. Data from 2011 show an increase in the number of Hispanic patients aged 25–34 that continues through the younger age groups (17). This population wave of younger Hispanics is just about to enter the age demographic that typically pursues reproductive services. Interestingly, approximately 60% of Hispanic Americans speak Spanish regularly (17). Further studies should examine whether the quality and the increase of Spanish-translated content on REI practice websites affect the use of reproductive services by the Hispanic patient population in the United States. Acknowledgments: The authors thank Carol Thompson and Ximin Li from the Johns Hopkins Biostatistics Center for their assistance with some of the statistical analyses, with support from the National Center for Research Resources and the National Center for Advancing Translational Sciences of the National Institutes of Health grant no. 1UL1TR001079. We also thank Caroline Moreau, M.D., Ph.D., M.P.H. from the Department of Population, Family and Reproductive Health at the Johns Hopkins Bloomberg School of Public Health for revisions and suggestions on the preliminary version of this manuscript.
REFERENCES 1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
United States Census Bureau. The Hispanic population 2010. Available from: http://www.census.gov/prod/cen2010/briefs/c2010br-04.pdf. Accessed March 25, 2014. Reddy S. Latinos fuel growth in the decade. Wall Street Journal, March 25, 2011. Available from: http://online.wsj.com/article/SB10001424052748704 604704576220603247344790.html. Accessed March 25, 2014. Waldman M. Terry College of Business, University of Georgia. Hispanic consumer market in the US is larger than the entire economies of all but 13 countries in the world. (Press Release). May 1, 2012. Available from: http:// www.terry.uga.edu/news/releases/hispanic-consumer-market-in-the-u.s.is-larger-than-the-entire-economies-of. Accessed March 25, 2014. eMarketer. Among Hispanics, who's leading digital adoption trends? Available from: http://www.emarketer.com/Article/Among-Hispanics-WhosLeading-Digital-Adoption-Trends/1009755#R0MGji8PsLZAlyDl.99. Accessed March 25, 2014. Feinberg EC, Larsen FW, Wah RM, Alvero RJ, Armstrong AY. Economics may not explain Hispanic underutilization of assisted reproductive technology services. Fertil Steril 2007;88:1439–41. McCarthy-Keith DM, Schisterman EF, Robinson RD, O'Leary K, Lucidi RS, Armstrong AY. Will decreasing assisted reproduction technology costs improve utilization and outcomes among minority women? Fertil Steril 2010;94:2587–9. Nachtigall RD, Castrillo M, Shah N, Turner D, Harrington J, Jackson R. The challenge of providing infertility services to a low-income immigrant Latino population. Fertil Steril 2009;92:116–23. Centers for Disease Control and Prevention. 2011 Assisted Reproductive Technology National Summary Report. Available from: http://www.cdc.gov/ art/ART2011/NationalSummary_index.htm. Accessed March 25, 2014. Motel S, Patten E. Pew Research Hispanic Trends Project. Characteristics of the 60 largest metropolitan areas by Hispanic population. September 19, 2012. Available from: http://www.pewhispanic.org/2012/09/19/character istics-of-the-60-largest-metropolitan-areas-by-hispanic-population/. Accessed March 25, 2014. Schiebinger L, Klinge I, Sanchez de Madariaga I, Schraudner M, Stefanick M, eds. Gendered innovations in science, health and medicine, engineering, and environment. 2011–2013. Available from: http://genderedinnovations. stanford.edu. Accessed March 25, 2014. Sharif I, Tse J. Accuracy of computer-generated, Spanish-language medicine labels. Pediatrics 2010;125:960–5. 1375
ORIGINAL ARTICLE: ENVIRONMENT AND EPIDEMIOLOGY 12.
13.
14.
Livingston G, Minushkin S, Cohn D. Hispanics and health care in the United States: access, information and knowledge, August 2008. Available from: http://pewhispanic.org/files/reports/91.pdf. Accessed March 25, 2014. Homberger JC, Gibson CD, Wood W, Dequeldre C, Corso I, Palla B, et al. Eliminating language barriers for non-English speaking patients. Med Care 1996;34:845–56. Lee LJ, Batal HA, Maselli JH, Kutner JS. Effect of Spanish interpretation method on patient satisfaction in an urban walk-in clinic. J Gen Intern Med 2002;17:641–5.
1376
15. 16.
17.
Omurtag K, Turek P. Incorporating social media into practice: a blueprint for reproductive health providers. Clin Obstet Gynecol 2013;56:463–70. Brenner J, Smith A. 72% of Online adults are social networking site users. Available from: http://pewinternet.org/Reports/2013/social-networkingsites.aspx. Accessed March 25, 2014. The Nielsen Company. The state of Hispanic consumer: Hispanic market imperative report, 2012. Available from: http://www.nielsen.com/content/ dam/corporate/us/en/reports-downloads/2012-Reports/State-of-the-His panic-Consumer.pdf. Accessed March 25, 2014.
VOL. 102 NO. 5 / NOVEMBER 2014
Fertility and Sterility®
SUPPLEMENTAL TABLE 1 Examples of automatic translation errors on REI practice websites. Original English Although he had many opportunities to go elsewhere, when I learned . (describing how a partner was recruited for the practice). When your body is ready for treatment, so are we. Doctors and nurses took away any anxiety (patient's testimonial). An oral medication, clomiphene citrate, is taken on days five to nine of the menstrual cycle (description of clomiphene protocol). Talk to an (practice name here) Doctor today! Free second opinion When we arrived we were told about the free IVF drawing (patient's testimonial). We're (location of the practice here)'s trusted full-service fertility clinic. Fertility News
Automatic Spanish translation
Actual translation meaning
Aunque no tenía muchas oportunidades de ir a otro lugar, cuando me entere.
Although he had not many other opportunities to go elsewhere.
Cuando su cuerpo esta listo para su tratamiento, así somos nosotros. n Los medicos y enfermeras se llevaron ningu tipo de ansiedad. Un medicamento oral citrato de clomifeno, se toma el día ocho y cincuenta y cinco del ciclo menstrual.
When your body is ready for treatment, that's the way we are. Doctors and nurses took away none kind of anxiety. An oral medication clomiphene citrate, is taken on the day eight and fifty-five of the cycle.
Hable con un medico (translated name of the practice here) hoy! n libre Segunda opinio Cuando llegamos nos dijeron que el dibujo a la FIV. Estamos clínica de fertilidad de servicio completo de confianza de (location of the practice here). Fertilidad de Noticias
Talk to a GUNS Doctor today! Released second opinion When we arrived we were told about a FIV layout. We are clinic of fertility of service full of confidence in (location of the practice here). Fertility of the news
Londra. Spanish language in REI websites. Fertil Steril 2014.
VOL. 102 NO. 5 / NOVEMBER 2014
1376.e1
Division of Reproductive Endocrinology and Infertility, Department of Gynecology and Obstetrics, Johns Hopkins School of Medicine, Lutherville, Maryland; b Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, Washington University School of Medicine, St. Louis, Missouri; and c Freelance Enterprise Resource Planning (ERP) Software Consultant, Association of Reproductive Managers, Baltimore, Maryland
Objective: To analyze the use of Spanish language translation on the websites of reproductive endocrinology and infertility (REI) practices in the context of evidence of underuse of infertility services by minority populations. Design: Cross-sectional survey of websites from REI practices. Setting: Not applicable. Patient(s): None. Intervention(s): None. Main Outcome Measure(s): Assessment of the relationship between having a Spanish-translated website and REI practice characteristics. Variables included concurrent use of social media, size of the practice, Spanish-speaking practitioner in the practice, being a private or a university-based practice, being in a mandated insurance state, and being in an area with different levels of percentage of Hispanic population, adjusted for annual income levels of the population. Result(s): Of the 376 REI practice websites analyzed, 101 (27%) offered at least some information in Spanish. We identified 97 Spanishspeaking practitioners at 71 REI practices. Having a Spanish-translated website was significantly associated with the practice's use of social media, having an international/out-of-town web page, and having a Spanish-speaking physician in the practice. The size of the practice, as measured in number of cycles reported per year, was not associated with having a translated website. In practices located in the top 60 metropolitan areas by Hispanic population, the odds of having a Spanish-translated website were only related to the percentage of Hispanic population after adjusting for state-mandated insurance and average annual income level of the Hispanic population. Sixty-six of the websites with Spanish-translated content had been automatically translated. An additional eight websites were partially translated automatically. Conclusion(s): REI practices in metropolitan areas with a higher percentage of Hispanics were more likely to reach out to this minority population by translating their website content into Spanish. These practices were also more likely to use social media. Future studies are needed to determine whether the availability of Spanish language content on REI websites is associated with increased use of reproductive services by this minority population. (Fertil SterilÒ Use your smartphone 2014;102:1371–6. Ó2014 by American Society for Reproductive Medicine.) to scan this QR code Key Words: Minority health, practice management, health communication, consumer health and connect to the information, Hispanic Americans, translating Discuss: You can discuss this article with its authors and with other ASRM members at http:// fertstertforum.com/londral-spanish-language-rei-practice-websites/
T
he U.S. Hispanic population surged by 43% between 2000 and 2010, rising to 50.5 million in 2010 (1). The U.S. Census Bureau projects that the economic power of
Hispanics in the United States will rise to more than $1.2 trillion in 2012 (2), larger than the entire economies of all but 13 countries in the world (3). This demographic growth is paralleled by
Received April 9, 2014; revised and accepted July 25, 2014; published online September 11, 2014. L.C.L. has nothing to disclose. K.J.T. has nothing to disclose. K.R.O. has nothing to disclose. M.B.D. has nothing to disclose. Reprint requests: Laura C. Londra, M.D., Division of Reproductive Endocrinology and Infertility, Department of Gynecology and Obstetrics, Johns Hopkins School of Medicine, 10751 Falls Road, Suite 280, Lutherville, Maryland 21093 (E-mail: [email protected]). Fertility and Sterility® Vol. 102, No. 5, November 2014 0015-0282/$36.00 Copyright ©2014 American Society for Reproductive Medicine, Published by Elsevier Inc. http://dx.doi.org/10.1016/j.fertnstert.2014.07.1243 VOL. 102 NO. 5 / NOVEMBER 2014
discussion forum for this article now.*
* Download a free QR code scanner by searching for “QR scanner” in your smartphone’s app store or app marketplace.
the increasing rate at which the Hispanic population is accessing medical information online (4). Yet studies have shown that utilization of assisted reproductive technology (ART) services by Hispanic women has not increased, even when economic barriers are minimized (5). In a study of military personnel, despite enhanced access to reproductive services, Hispanic patients did not use ART to the same extent as other ethnic groups (6). Other factors, such as social and cultural influences, 1371
ORIGINAL ARTICLE: ENVIRONMENT AND EPIDEMIOLOGY may explain the significantly lower use of reproductive services compared with the white and African American populations (5). After accounting for cost of treatment, Hispanics indicated that they were less inclined to seek out infertility treatments because of language or cultural barriers, lack of continuity with providers, difficulty scheduling visits and procedures, and a perceived decreased availability of services (7). There is a need to identify and remove barriers faced by Hispanic couples who wish to access information about reproductive services and ART. We hypothesized that communication of reproductive health services to Hispanic populations in the United States is limited by an overall lack in quantity of Spanish language health information delivered via reproductive endocrinology and infertility (REI) practice websites. In addition, we hypothesized that practices that provide Spanish language translations would also have similar features, such as Spanish-speaking physicians. We tested these hypotheses by analyzing the websites of REI practices in the United States for their use of Spanish language translation and its correlation with various practice characteristics and the Hispanic demographics of the practice locations.
MATERIALS AND METHODS Data Collection
The Johns Hopkins University School of Medicine Institutional Board Review (IRB) determined that this type of study is exempt from IRB review. From September 2013 to October 2013, we analyzed the websites of all practices that report to the Society for Assisted Reproductive Technology (SART), which represents approximately 88% of all REI practices in the United States (8). The SART website was used to identify all infertility clinics listed under the ‘‘Find a Clinic’’ tab, which is part of a public domain of the SART website. Using the name and address information from the SART website, Google and Bing searches were used to find the clinic websites. We collected data on practice characteristics based on the information available on the practices' websites, including the availability of Spanish-speaking physicians and their reported level of fluency, use of social media, whether the practice setting was private or university based, existence of a web page directed to international and outof-town patients, additional languages translated, and whether or not the practice was located in a state with mandated insurance coverage for reproductive services. The size of the practice was estimated by using the number of IVF cycles reported per year; these data were then used to perform logistic regression to estimate the odds of having a translated website for each increment in 100 reported cycles per year. For Spanish-translated websites, we distinguished between automatic translation and human translation.
Statistical Analysis Presence of Spanish translation on the websites was used as the dependent variable. The availability of Spanishspeaking staff within the REI practice, the use of social media, 1372
private or academic practice setting, and location in a state with mandated insurance coverage for reproductive services were treated as independent variables in univariate logistic regression analyses, accounting for within-state correlations of characteristics. The size of the practice was evaluated by using the number of reported IVF cycles per year as a surrogate for the number of patients treated and the wealth of the practice. Logistic regression was conducted using increments of 100 cycles per year, clustered on state. We then estimated the odds of having a translated website with each increment in the number of annual cycles per practice. Additionally, the top 60 U.S. metropolitan areas by Hispanic percentage and total population (9) were correlated with the information obtained from the REI practice websites as follows: we noted all of the practices that were located within the 60 largest metro areas by Hispanic population. Then the percentage of Hispanic population from these metropolitan areas and their average annual income levels were categorized separately and clustered by state. Categories of percentage of Hispanic population were as follows: 0–10%, >10%–30%, and >30%. Average annual income categories were (in U.S. dollars) 30,000 to 40,000, >40,000 to 50,000, >50,000 to 60,000, and >60,000 to 70,000. These independent variables were used to conduct logistic regression analysis, accounting for within-state correlation of website characteristics, to identify factors associated with the presence of REI practices that have a translated website, adjusting for category of percentage of Hispanic population, category of annual income, and location in a state with mandated insurance for reproductive services. Analyses were performed using STATA (v 11, Statacorp); P< .05 was considered statistically significant.
RESULTS A total of 376 of 396 SART member practices were included in the analysis. Twenty clinics in the SART registry were excluded either because they did not have a website or because the website was a duplicate with only the practice location and/or URL information being different. Upon review of the 376 websites, 104 (28%) contained content with foreign language translation. The overwhelming majority (97%) of translated websites provided a Spanish language translation. Only three websites did not offer Spanish; those sites were translated into Chinese and Japanese. Among the 101 REI websites that offered Spanish translation, 27 used a human translator and 66 had installed a widget that allows users to automatically translate the website material into the target language. Eight practices used a combination of both, using a launch or splash page with the essential details translated into Spanish. The combined option provided potential patients with a localized introduction to the practice, plus access to the entire site (note: localization is defined as translation and adaptation of content for a particular audience). Nineteen of the 27 practices that employed professional or human translators provided a complete or nearly complete translation of their website content. Results from logistic regression analyses evaluating the association of Spanish language translation with various VOL. 102 NO. 5 / NOVEMBER 2014
Fertility and Sterility®
TABLE 1 Association between practice characteristics and having a Spanish-translated website. Translated website, No translated website, n [ 101 (%) n [ 275, (%)
Practice characteristica In a state with mandated insurance for infertility services Uses social media Includes a Spanish-speaking physician Has a web page directed to international/out-of-state patients University based Private practice a b
63 (62.3) 80 (79.2) 32 (31.7) 41 (40.6) 20 (19.8) 81 (80.2)
138 (50.2) 165 (60) 39 (14) 32 (11.6) 58 (21.1) 212 (77.1)
OR 1.64 2.50 2.80 5.18 0.92 1.20
95% Confidence interval P valueb 0.74–3.62 1.71–3.76 1.72–4.56 2.9–8.97 0.52–1.63 0.68–2.10
.215 < .001 < .001 < .001 .521 .517
Reference category for each predictor is not having the characteristic. Logistic regression adjusting for within-state correlation of website characteristics.
Londra. Spanish language in REI websites. Fertil Steril 2014.
independent variables for all websites and for those in the top 60 metropolitan areas in the United States are shown in Tables 1 and 2, respectively. Logistic regression on number of annual cycles reported per practice using increments of 100 cycles per year, clustered on state, did not show a significant association with having a translated website (OR, 1.05; 95% CI, 0.99–1.11; P¼ .076). Upon examination of the practices located in the 60 largest metropolitan areas in the United States (9), in general, the percentage of the local Hispanic population was associated with practices having Spanish translation of their websites (Fig. 1). Exceptions did exist: 85% of practices in the Boston metropolitan area had websites with Spanish translation, despite Boston having only a 10% Hispanic population. Annual income categories of the Hispanic population were analyzed using logistic regression, and the income category 4 (>60,000–70,000 U.S. dollars) was considered the reference category. There was a statistically significant difference across the three highest income categories (>40,000 U.S. dollars per year) in Hispanic population income categories. The annual income category 3 (>50,000–60,000 U.S. dollars) and annual income category 2 (>40,000–50,000 U.S. dollars) were associated with increased odds of availability of REI websites with Spanish translation in the area (OR, 7.20; 95% CI, 1.89–27.55; P¼ .004; and OR, 4.95; 95% CI, 1.17–20.99, respectively), adjusted by state where the practice is located. As shown in Table 1, 71 physicians indicated on their practice websites that they had some degree of Spanish proficiency. Physicians used a variety of terms on their web
TABLE 2 Likelihood of having a Spanish-translated website by categories of percentage of Hispanic population in the 60 largest metropolitan areas in the United States.a OR
95% Confidence interval
P value
Reference 4.26 7.43
– 1.55–11.7 2.83–19.48
– .005 < .001
Hispanic population (%) 0–10 11–30 >30
a Logistic regression analysis was used to calculate OR, adjusting for categories of average annual income level of the Hispanic population and state-mandated insurance, and accounting for within-state correlation of responses.
Londra. Spanish language in REI websites. Fertil Steril 2014.
VOL. 102 NO. 5 / NOVEMBER 2014
page to describe their proficiency in Spanish. The selfdescribed capability was as follows: ‘‘Spanish’’ or ‘‘Spanish spoken’’ (32%); ‘‘Fluent in Spanish’’ (30%); ‘‘Graduate of a university in a Spanish-speaking country’’ (13%), usually as part of a physician profile; ‘‘proficient/conversant in Spanish’’ (4%); bilingual or multilingual (4%); and other (16%). Having a Spanish-speaking physician on staff was also found to be associated with practices translating their website into Spanish. Our analysis also found that 245 REI practices (65%) engaged in social media through their websites. Of those, 97% of them displayed links to Facebook, followed by Twitter (75%), YouTube (43%), Googleþ (28%), LinkedIn (19%), RSS (19%), and Pinterest (14%). Practices engaging in social media had 2.5-times higher odds of having Spanish-translated content than websites not using social media (Table 1). Seventy-three REI practices made an effort to serve patients outside their local area by offering an ‘‘international or out-of-town’’ web page. Interestingly, almost half of these websites offered no translation to a foreign language (Table 1). Most of these practices included on these pages steps on how to make the IVF experience a successful one, providing new patient forms and listing the necessary medications and steps to take before arriving at the office. A few practices used their websites to promote overseas recruiting trips to conduct seminars for prospective patients.
DISCUSSION We found that just over one-quarter of REI practices sampled in the United States had Spanish-translated content on their practice websites (Fig. 2). Generally, practices located in areas with a large Hispanic population were more likely to have Spanish-translated website content, suggesting that these practices recognize the need for outreach to the growing number of Hispanic health consumers in the United States. The estimated annual income levels of the Hispanic population appeared to have an influence only in the middle categories (annual income >40,000–60,000 dollars per year), likely reflecting the fact that patients with a higher income might be more likely to be more educated and therefore less sensitive to the lack of a translated website when looking for infertility care services. 1373
ORIGINAL ARTICLE: ENVIRONMENT AND EPIDEMIOLOGY
FIGURE 1
Percent of REI practices in U.S. metropolitan areas providing Spanish language translation on their websites, by Hispanic population. The total number of Hispanics in each city is represented by the size of each sphere. Placement of each sphere on the graph illustrates the relationship between the size of the Hispanic population and the percentage of practices that have Spanish-translated websites for each city. For example, 100% of REI practices in the Ft. Lauderdale area (which is 25% Hispanic) translated their websites into Spanish, compared with 85% of REI practices in Boston that translated their websites into Spanish, despite having a population that is only 10% Hispanic. Owing to space constraints, only the 10 cities most and least likely to translate their website are included in the figure (in addition, New York, Los Angeles, and Houston were selected for comparison purposes). Demographic data are from the Pew Research reference (9). Londra. Spanish language in REI websites. Fertil Steril 2014.
Machine translation is cost-effective and does a remarkable job of producing instant translations in an extensive array of languages. However, machine translators often
struggle with nuances of human language (Supplemental Table 1). Machine translation also ignores the fact that U.S. Hispanics come from 22 different countries, with
FIGURE 2
Most frequently offered language translations on U.S. REI practice websites. Each unit represents a link to a individual website translation, except for the Google translate link, which supports 80þ languages, including Spanish. Londra. Spanish language in REI websites. Fertil Steril 2014.
1374
VOL. 102 NO. 5 / NOVEMBER 2014
Fertility and Sterility® various language preferences and degrees of U.S. acculturation. A study on machine translation and gender found that source texts used with Google's translation algorithms lead to unintentional sexism, such as overly assuming that doctors are men and teachers are women (10). The issue of translation in health care was exemplified in 2009, when New York pharmacies were required to provide multilingual medicine labeling; the effort to comply revealed the difficulty in successfully incorporating automatic translation into medical practice. A subsequent study in 2010 showed that machine translations were associated with up to 50% of prescription translations with serious errors (11). Therefore, potential inaccuracies in machine translations may not only misinform patients but also damage the image of the practice. While physicians recognize the need to reach out to Hispanic populations through Spanish translation of their practice websites, the language barrier may continue to pose a challenge to providing care to Hispanic patients. A study by the Pew Hispanic Center found that a mismatch between Hispanic patients' primary language and the language spoken during their doctor appointments lowered the patientreported satisfaction with their care. Up to 30% of Spanish speakers whose appointments were usually conducted in English rated their care as fair to poor, compared with only 19% percent of those whose appointments are conducted in Spanish (12). For those physicians who cannot adequately explore the complexities and nuances of patient interactions in Spanish, an alternative may be to make use of the skills of a trained medical interpreter or translator. Use of remote interpreting services (such as telephone interpreters) also has proven to be very successful (13, 14). We found that two-thirds of REI practice websites engaged in social media, dramatically higher than the less than a third of practices reported just 2 years ago. The use of social media to connect with patients, regardless of ethnicity, appears to be increasing in importance among medical practices. Previous research (15) has shown a positive association between clinic size and the use of social media; we did not find an association between a translated website and clinic size, but we did find a correlation between the use of social media and having a translated website. A recent Pew Report found that ‘‘Hispanics are the most active group on social media networking sites, with 80 percent of Hispanic adults in the U.S. using social media, followed by African Americans (75 percent) and Caucasians (70 percent)’’ (16). It remains unclear whether Hispanics use social media in the context of their health care. Interestingly, practices that had a translated website were also more likely to use social media and to display an out-of-town page for potential patients from other states or other countries. These findings suggest that REI practices in the United States are increasingly open to interact with and serve patients beyond the local residents regardless of clinic size as estimated by number of cycles per year. Overall, more than half of all REI practices neither translate their website nor identify a Spanish-speaking physician on staff. Population demographics may explain why some practices concentrate on communicating with the nonVOL. 102 NO. 5 / NOVEMBER 2014
Hispanic white population. Data from 2011 show an increase in the number of Hispanic patients aged 25–34 that continues through the younger age groups (17). This population wave of younger Hispanics is just about to enter the age demographic that typically pursues reproductive services. Interestingly, approximately 60% of Hispanic Americans speak Spanish regularly (17). Further studies should examine whether the quality and the increase of Spanish-translated content on REI practice websites affect the use of reproductive services by the Hispanic patient population in the United States. Acknowledgments: The authors thank Carol Thompson and Ximin Li from the Johns Hopkins Biostatistics Center for their assistance with some of the statistical analyses, with support from the National Center for Research Resources and the National Center for Advancing Translational Sciences of the National Institutes of Health grant no. 1UL1TR001079. We also thank Caroline Moreau, M.D., Ph.D., M.P.H. from the Department of Population, Family and Reproductive Health at the Johns Hopkins Bloomberg School of Public Health for revisions and suggestions on the preliminary version of this manuscript.
REFERENCES 1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
United States Census Bureau. The Hispanic population 2010. Available from: http://www.census.gov/prod/cen2010/briefs/c2010br-04.pdf. Accessed March 25, 2014. Reddy S. Latinos fuel growth in the decade. Wall Street Journal, March 25, 2011. Available from: http://online.wsj.com/article/SB10001424052748704 604704576220603247344790.html. Accessed March 25, 2014. Waldman M. Terry College of Business, University of Georgia. Hispanic consumer market in the US is larger than the entire economies of all but 13 countries in the world. (Press Release). May 1, 2012. Available from: http:// www.terry.uga.edu/news/releases/hispanic-consumer-market-in-the-u.s.is-larger-than-the-entire-economies-of. Accessed March 25, 2014. eMarketer. Among Hispanics, who's leading digital adoption trends? Available from: http://www.emarketer.com/Article/Among-Hispanics-WhosLeading-Digital-Adoption-Trends/1009755#R0MGji8PsLZAlyDl.99. Accessed March 25, 2014. Feinberg EC, Larsen FW, Wah RM, Alvero RJ, Armstrong AY. Economics may not explain Hispanic underutilization of assisted reproductive technology services. Fertil Steril 2007;88:1439–41. McCarthy-Keith DM, Schisterman EF, Robinson RD, O'Leary K, Lucidi RS, Armstrong AY. Will decreasing assisted reproduction technology costs improve utilization and outcomes among minority women? Fertil Steril 2010;94:2587–9. Nachtigall RD, Castrillo M, Shah N, Turner D, Harrington J, Jackson R. The challenge of providing infertility services to a low-income immigrant Latino population. Fertil Steril 2009;92:116–23. Centers for Disease Control and Prevention. 2011 Assisted Reproductive Technology National Summary Report. Available from: http://www.cdc.gov/ art/ART2011/NationalSummary_index.htm. Accessed March 25, 2014. Motel S, Patten E. Pew Research Hispanic Trends Project. Characteristics of the 60 largest metropolitan areas by Hispanic population. September 19, 2012. Available from: http://www.pewhispanic.org/2012/09/19/character istics-of-the-60-largest-metropolitan-areas-by-hispanic-population/. Accessed March 25, 2014. Schiebinger L, Klinge I, Sanchez de Madariaga I, Schraudner M, Stefanick M, eds. Gendered innovations in science, health and medicine, engineering, and environment. 2011–2013. Available from: http://genderedinnovations. stanford.edu. Accessed March 25, 2014. Sharif I, Tse J. Accuracy of computer-generated, Spanish-language medicine labels. Pediatrics 2010;125:960–5. 1375
ORIGINAL ARTICLE: ENVIRONMENT AND EPIDEMIOLOGY 12.
13.
14.
Livingston G, Minushkin S, Cohn D. Hispanics and health care in the United States: access, information and knowledge, August 2008. Available from: http://pewhispanic.org/files/reports/91.pdf. Accessed March 25, 2014. Homberger JC, Gibson CD, Wood W, Dequeldre C, Corso I, Palla B, et al. Eliminating language barriers for non-English speaking patients. Med Care 1996;34:845–56. Lee LJ, Batal HA, Maselli JH, Kutner JS. Effect of Spanish interpretation method on patient satisfaction in an urban walk-in clinic. J Gen Intern Med 2002;17:641–5.
1376
15. 16.
17.
Omurtag K, Turek P. Incorporating social media into practice: a blueprint for reproductive health providers. Clin Obstet Gynecol 2013;56:463–70. Brenner J, Smith A. 72% of Online adults are social networking site users. Available from: http://pewinternet.org/Reports/2013/social-networkingsites.aspx. Accessed March 25, 2014. The Nielsen Company. The state of Hispanic consumer: Hispanic market imperative report, 2012. Available from: http://www.nielsen.com/content/ dam/corporate/us/en/reports-downloads/2012-Reports/State-of-the-His panic-Consumer.pdf. Accessed March 25, 2014.
VOL. 102 NO. 5 / NOVEMBER 2014
Fertility and Sterility®
SUPPLEMENTAL TABLE 1 Examples of automatic translation errors on REI practice websites. Original English Although he had many opportunities to go elsewhere, when I learned . (describing how a partner was recruited for the practice). When your body is ready for treatment, so are we. Doctors and nurses took away any anxiety (patient's testimonial). An oral medication, clomiphene citrate, is taken on days five to nine of the menstrual cycle (description of clomiphene protocol). Talk to an (practice name here) Doctor today! Free second opinion When we arrived we were told about the free IVF drawing (patient's testimonial). We're (location of the practice here)'s trusted full-service fertility clinic. Fertility News
Automatic Spanish translation
Actual translation meaning
Aunque no tenía muchas oportunidades de ir a otro lugar, cuando me entere.
Although he had not many other opportunities to go elsewhere.
Cuando su cuerpo esta listo para su tratamiento, así somos nosotros. n Los medicos y enfermeras se llevaron ningu tipo de ansiedad. Un medicamento oral citrato de clomifeno, se toma el día ocho y cincuenta y cinco del ciclo menstrual.
When your body is ready for treatment, that's the way we are. Doctors and nurses took away none kind of anxiety. An oral medication clomiphene citrate, is taken on the day eight and fifty-five of the cycle.
Hable con un medico (translated name of the practice here) hoy! n libre Segunda opinio Cuando llegamos nos dijeron que el dibujo a la FIV. Estamos clínica de fertilidad de servicio completo de confianza de (location of the practice here). Fertilidad de Noticias
Talk to a GUNS Doctor today! Released second opinion When we arrived we were told about a FIV layout. We are clinic of fertility of service full of confidence in (location of the practice here). Fertility of the news
Londra. Spanish language in REI websites. Fertil Steril 2014.
VOL. 102 NO. 5 / NOVEMBER 2014
1376.e1