Teaching Spanish to Emergency Medicine Residents

~~ ~~ ~~ ~

32 ACADEMIC EMERGENCY MEDICINE JAN 1995 VOL 2/NO 1

Teaching Spanish to Emergency Medicine Residents

Diana Prince, BS, Marc Nelson, MD, PhD

I ABSTRACT

Objective: To determine the effects of teaching medical Spanish to eight PGYl emergency medicine residents on their subsequent interactions with Spanish-speaking patients.

Methods: Eight PGY 1 residents completed a 45-hour medical Spanish course administered during their first residency month. Thirty-four subsequent physician-patient interactions by these residents were audiotaped over a six-month period at a suburban teaching ED. The tapes were transcribed and analyzed for errors by a professional medical Spanish interpreter and a native Spanish speaker.

Results: Minor errors (e.g., technically incorrect grammar or vocabulary with generally appropriate patient understanding) were found in more than half of the interactions and major errors (e.g., misunderstanding duration of symptoms, misunderstanding of vocabulary) were found in 14% of the interactions. In addition, although the course was designed to supplement, not replace, professional interpreters, the residents called for an interpreter only 46% of the time.

Conclusion: Although medical language courses may be a useful ad- junct to interpreters, they are not designed to replace them. Signif- icant errors may occur when participants in such courses assume their knowledge is sufficient to obtain a good history, give patient release instructions, and provide medical care in general without an interpreter present.

Acad. Emerg. Med. 1995; 2:32-37.

I Data from the 1990 U.S. Census indicate that the Hispanic population has grown rapidly over the past ten years and currently rep- resents 9.0% of the U.S. population.' In certain Southwestern states, such as California, the Hispanic population comprises over 25% of the total population and is expected to increase to 30% by the year 2000.* A significant number of Hispanics speak little or no English. Preliminary work at Stanford University Hospital's E D showed that 12% of the patients were monolingual Spanish speakers who required an interpreter to communicate with medical personnel. In light of this information, we studied the effects of teaching a limited amount of medically oriented Spanish to eight PGYl residents during their first month of orientation in the Stanford University/Kaiser Emer- gency Medicine (EM) Residency program. We sought to determine whether such a course would be beneficial to patients and residents by improving patient rapport, increasing communication, and avoid- ing errors inherent in bilingual interactions.

Teaching Spanish to EM Residents, Prince, Nelson 33

I METHODS . . . . . . , . . . . , . . . . , . . . , . . , . . . . . . . . . . . . . . . . . . , , , . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . , , . . . . . . . . . . . . , . . . . . . . .

Study Design We conducted a prospective observational study of

the effects of a 45-hour introductory medical Spanish course on physician- patient communication (especially translation errors) of English-speaking EM residents during interaction with Spanish-speaking patients.

Population Eight PGYl EM residents in the Stanford/Kaiser

EM program completed the Spanish course and were subsequently audiotaped while interviewing Spanish- speaking patients. Of the eight residents who participated, four had no prior knowledge of Spanish, two had very limited Spanish exposure consisting of high- school courses or travel experience, one had a quarter- year of college Spanish, and another had three years of college Spanish courses.

Audiotaping was conducted at the Stanford Hospital ED, a suburban E D located in a major academic medical center. Thirty-four interactions between the PGY 1 resident physician and the patient (or family if the patient was a minor) were audiotaped over a six-month period between October 1991 and May 1992. The patients ranged in age from 2 months to 85 years. There were 14 pediatric patients (average age 2 years) and 20 adult patients (average age 40 years). Overall, there were 22 female and 12 male patients. The educational levels of those interviewed (patient, or parents in the case of a minor) ranged from none to some college: 12 had no formal education, nine had some or all of ele- mentary school, seven had some high school, three graduated from high school, and three had some college. Twenty-nine of the patients were from Mexico, three were from El Salvador, one was from Guatemala, and one was from Colombia.

The participants were selected on the basis of their need to interact entirely in Spanish, and all gave verbal consent to participate in the study. Spanish-speaking patients who were too ill to give consent were not included in the study. All patients and parents who were asked to participate consented to do so. The study was approved by the Human Subjects Committee at Stan- ford University.

Course Description The eight PGYl residents were required to attend

a 45-hour medical Spanish course during their first month of residency. The class was held for ten hours per week in two half-hour sessions for four and a half weeks. All the residents attended the course as required. The stated goals of the course were to have the residents: I) learn

conversational Spanish with an emphasis on medical communications skills; 2) establish positive physician- patient relationships; and 3) obtain a medical history from and conduct a physical examination on Spanish- speaking patients. Topics covered in the course included basic pronunciation, vocabulary, grammar, and role-playing pertinent to the types of medical situations encountered (pediatric patients, obstetric patients).

Upon completion of the course, the residents’ per- formances were evaluated by the class instructor through both written examination and oral interview. Through- out the course, the residents were reminded that the purpose of the instruction was to enhance, not replace, the use of professional interpreters.

Measurements A convenience sample of 34 resident physician-pa-

tient interactions was obtained at the ED of the academic medical center. All the residents were audiotaped two to six times each, with the mean being five interviews per resident. The entire patient visit was ob- served by one of the authors (DP), who took notes on both verbal and nonverbal interactions. At the end of each visit, the physicians filled out a short questionnaire asking about their perceptions of the interview using a Likert-type scale format. Questions included how the physicians thought the patients felt about their speaking in Spanish and how well the physicians thought the patients understood them. In addition, each patient was interviewed in Spanish regarding the same issues.

The audiotapes were transcribed by a professional medical Spanish interpreter and a native Spanish speaker. The transcriptions were then analyzed by the same team. Each error identified was agreed upon by both inves- tigators and was discussed as to its degree of significance. The tapes were analyzed for errors in both speaking and understanding Spanish. Errors were classified as either minor or major.

Minor errors in speaking Spanish included gram- matical or vocabulary mistakes in which the general meaning of the response or question was understood by the patient. A minor error in understanding Spanish was defined as one in which the physician generally understood the gist of the question or response by the patient. Minor errors were often followed with additional questions or responses between the physician and the patient. This subsequent exchange helped to clarify the meaning.

Major errors were classified as those questions or responses that were not understood by the patient or physician and could not be clarified by additional con- versation between the two. Additionally, each major mistake was analyzed to see whether it resulted in an error in delivery of health care.


The significance of each error regarding health care delivery was subsequently determined by an attending physician at the Stanford ED. Medically significant mistakes were defined as a misdiagnosis, a prolonged delay in diagnosis, or the ordering of unnecessary additional tests. Medical significance also was based on the patient’s understanding of both diagnosis and treatment as determined by the postvisit interview and analysis by a professional interpreter.

Data Analysis The number of participants/sample size was small

and prohibited extensive quantitative analysis. How- ever, the goal of our study was not to provide quantitative analysis of translation errors, but to provide a descriptive, qualitative analysis. Therefore, only simple descriptive statistics are reported.

I RESULTS

Of the 34 taped interviews, the interpreters determined that six contained too little Spanish to ascertain the resident’s abilities. The 28 remaining interviews were classified by the type of medical probtem presented. There were ten internal medicine complaints, ten pediatric complaints, five surgical complaints, two psychiatric complaints, and one obstetric/gynecologic complaint. Interactions were classified by the type of problem, because the language skill level required when evalu- ating a patient with a sprained ankle differs from that required for the assessment of a patient with a psychiatric complaint. The former requires only limited language proficiency; the latter, quite a bit more.

In four of the interviews (14%), no error in speaking or understanding Spanish was made. In 12 of the interviews (43%), one or a few (< 4) minor errors were made. In six of the interviews (21%), more than four minor errors were made. In four of the interviews (14%), one major error was made, and in two of the interviews (7%), two major errors were made.

Examples of minor errors included misuse of the Spanish verbs set and esrur (both mean “to be,” but with different applications), misuse of masculine or feminine nouns or articles, mistakes in proper-person forms of verbs, mixing of Spanish and English, mistakes in present vs past tenses, and mistakes due to speaking and understanding a limited Spanish vocabulary.

Although the minor errors did not result in loss of understanding the basic message between physician and patient in the interpreter’s impressions, the misuse of serlesrar did result in an insulting statement’s being made by a physician to a patient. Specifically, the physician asked the patient the following question in Spanish: iQue otras personas de su familia son malos tambikn

ahorita? The physician meant to ask “are there others sick in the family right now”; however, this statement actually means “are there any other bad people in your family?”

Examples of major errors include misunderstanding of a patient’s symptoms or duration of symptoms (2), relying on incorrect information given by an unqualified interpreter (3), not recognizing that a particular question was not answered ( l ) , misunderstanding a patient’s vocabulary ( l ) , interrupting a patient in the middle of a sentence (l), asking a question in the wrong tense ( l ) , and falsely assuming a patient was intoxicated (1). A major mistake occurred when a physician meant to ask a patient whether she had been drinking the pre- ceding night but instead asked her whether she was planning to drink that night. Another major error occurred when a physician asked a patient whether the two sides of her abdomerl felt the same. The patient answered that one side was better than the other but the friend interpreting said that both sides were all right. All major errors were made by residents who had had no knowledge of Spanish prior to the study.

At the beginning of the study, the residents said they anticipated using Spanish an average of eight times per week. At the end of the study, they estimated their actual use at 13 times per week. (The residents see approximately 100 patients per week.) All indicated both before and after the study that they found the training to be valuable to them as physicians. Further- more, all perceived their Latino patients to be appreciative of their efforts, which was confirmed by the patients who participated in the study. They indicated they were appreciative of the physicians’ efforts to communicate with them in Spanish.

1 DISCUSSION

Research suggests that patients whose native language is not English are more satisfied with medical interactions in their native language than with those in English, and they are more likely to comply with treatment instructions and return for follow-up care.3 In fact, all of the patients in our study appreciated the efforts of the physicians to speak Spanish. As one patient com- mented, “It’s a relief to finally have a doctor who tries to understand what I have to say.”

In an ideal situation all EDs would have trained professional interpreters available for non-English- speaking patients. Unfortunately, this is rarely possible. Often, paid interpreters must serve an entire hospital, usually for limited hours, and rareiy is an interpreter immediately available 24 hours a day. At our institu- tion, the few interpreters who are employed are very busy, often resulting in increased waiting time, dissat- isfaction, and frustration for Spanish-speaking patients.

Teaching Spanish to EM Residents, Prince, Nelson 35

Of the 34 physician-patient interactions of this study, 13 (46%) used a professional interpreter. The interpreter was used for seven internal medicine complaints, four surgical complaints, one pediatric complaint, and one psychiatric complaint. Eight of the interactions had the interpreter present during the initial history, and five requested one after the initial history. For these 34 interactions, seven of the eight residents requested an interpreter at least once, and four of the seven more than once. No major error was made during the interviews where a professional interpreter was present. The residents who had prior knowledge of Spanish called an interpreter just as frequently as did those who did not have any previous Spanish experience. The one resident who did not use an interpreter during the study had three years of college-level Spanish in addition to the required medical Spanish course.

The delay in obtaining an interpreter is especially problematic in the ED, where it is often critical to obtain information from a patient within a matter of min- utes after the patient’s a r r i ~ a l . ~ However, due to the lack of critically ill or injured patients in our study, we were not able to study this problem.

An assumption commonly is made that any bilingual person can adequately serve as a medical interpreter.s This incorrect assumption often leads to the use of friends, family, or miscellaneous medical personnel to act as interpreters. This practice can lead to compiex problems, including incorrect translation, distortion of information, exaggeration or minimization of symptoms, breach of confidentiality, and a variety of other issues attributed to a lack of understanding of medical ter- minology and procedures by unqualified interpretem6 In our study, seven of the interviews (25%) relied on a friend or family member as an interpreter. These seven included five of the six interviews containing one or more major errors. One had no error and one had many minor errors. None of the interviews containing major errors had a professional interpreter called during any part of the interview. O n average, the residents reported using a friend or family member as an interpreter 20% of the time.

Although one possible solution to the lack of interpreters is to increase the number of bilingual health care providers through recruitment and training programs, attempts to increase the numbers of ethnic mi- norities have not been suc~essfu l .~ Another solution, though not necessarily easy to implement, would be to train health care providers to speak a second language. Unfortunately, we were unable to find many programs that have implemented such an approach. We have identified only two residency programs of this type. The most extensive training was found at the Cross-cultural Family Medicine Residency Program at the University of California at San Diego. This program focused on

both Latino and East Asian cultures. The residency training included 52 hours of medical Spanish instruction given one hour per week during the internship. The goal of the program was to have residents conduct patient interviews and physical examinations without the use of an interpreter.x Although the program claimed success in this goal, the outcome measures described did not fully define or substantiate the program’s claimed “success. ”

The other program was at the Texas Tech University EM residency program. This program incorporated a 45-hour medical Spanish course given as a three-hour class per week for 15 weeks. The goals of this program were to have residents become proficient in common patient interactions and to make them aware of when an interpreter was needed. This program was evaluated by student ability to role-play in Spanish with the course instructor. The residents also were given questionnaires regarding their use of Spanish with patients. The residents indicated that they believed they relied less on interpreters and that they made no significant medical error with their Spanish skills. Evaluation by the Span- ish instructor indicated that “most participants could take uncomplicated histories, conduct a physical exam and give common discharge instructions by the conclusion of the course.”y

Although both of the aforementioned programs claimed positive benefits from brief Spanish instruction, the outcome measures either were not explained or were subjective measures by program participants, either the Spanish instructors or the residents themselves. Be- cause of these concerns and the need to study residents in a “natural” setting, we analyzed audiotapes of real situations where E M residents used Spanish language skills with patients. We also studied whether problems might result from this very brief Spanish training. In addition to language misinterpretation, we also were concerned that residents who had skill in translating might be pulled away from their own work to help interpret for other medical personnel. This did not occur. Finally, we investigated whether limited proficiency would affect the use of professional interpreters. This was of particular interest to us because the course was designed to supplement the use of an interpreter; not replace it. Unfortunately, the residents did not al- ways call for an interpreter. Furthermore, when ques- tioned, each resident said that on average they called an interpreter only 63% of the time.

1 LIMITATIONS AND FUTURE QUESTIONS

There are limitations to this study. The number of interactions analyzed was relatively small and may not be representative of the patient situations encountered by

. . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . , , . , , , . , , , , , , . . . . , . . . , , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


other physicians. Both the physicians and the patients were aware of being taped for the study, and this fact may have introduced some bias into the interactions. The effect of an observer on the interaction also may have resulted in bias. However, we believe that taping the situation is as close to real life as possible and is a better outcome measure than is role-playing or resident self-assessment. Further, knowledge of the study could be expected to increase the number of times interpreters were requested and encourage the resident to interact optimally. Hence, the finding of major errors in 14% of the cases and the use of medical interpreters only 46% of the time probably represent best-case estimates.

One of the most interesting findings of the study was the high error rate when family members were used as interpreters. This is a common practice, particularly regarding those languages for which interpreters are limited. This area merits further study and is currently being investigated at our hospital. The role of professional interpreter services (e.g., the 24-hour service pro- vided by the AT&T Corporation) in these circum- stances also should be assessed.

This study focused on patients with nonurgent problems. In the absence of immediately available interpretation services, it would be interesting to note how lack of interpretation services affects patient care. Perhaps the focus of training should be on urgent problems and not those in which one has the luxury of waiting for an interpreter.

Perhaps most importantly, the medical impact of the translation errors warrants further evaluation. Fortu- nately, in this study, most of the major errors either were cleared up later by an interpreter or were significant only in the context of insulting a patient or por- traying insensitivity. Only one of the major errors was determined to have had a possible negative impact on a patient’s care. The error occurred during a physician’s attempt to warn a pediatric patient’s parents about the dangers of giving the child aspirin for a high fever. The interpreters determined that these instructions were clearly not understood by the parents. Fortunately, the parents had been giving the child acetaminophen for fever.

I CONCLUSION .........................................................................................................

The increasing number of Spanish-speaking patients in the United States mandates a response from health care providers seeking to provide high-quality patient care. Medical Spanish courses offer one solution. There are, however, potential problems with these courses. Res- idents may assume their Spanish is adequate and may no longer use a professional interpreter. In such interactions, we found a significant number of both major

and minor language errors. In addition, similar problems were found when using friends or family members to translate. Although these courses may be useful and are clearly appreciated by patients, they cannot and should not replace professional interpreters.

I REFERENCES

1.

2.

3.

4.

5.

6.

7.

8.

9.

...

Statistical Abstract of the United States, 1990 (population). Wash- ington, DC: Bureau of Census, 1990. Hayes-Bautista D, et al. No Longer a Minority: Latinos and Social Policy. Los Angeles: UCLA, 1992. Kline F. Acosta F, Austin W, et at. The misunderstood Spanish speaking patient. Am J Psychiatry. 1980; 137:1530-3. Weaver C, Sklar D. Diagnostic dilemmas and cultural diversity in emergency rooms. West J Med. 1980; 133:356-66. Vasques C. Javier R. The problem with interpreters: communi- cating with Spanish speaking patients. Hosp Community Psychia- try. 1989; 43(2):163-5. Petersdorf RG. Not a choice, an obligation. Acad Med. 1992;

Putsch R. Cross cultural communication: the special case of interpreters in health care. JAMA. 1985; 254:3344-8. Kristal L, Pennock P, Foote SM, et al. Cross cultural family medicine residency training. J Fam Pract. 1983; 17:683-7. Binder L, Nelson B. Smith D, et al. Development, implementation and evaluation of a medical Spanish curriculum for an emergency medicine training program. J Emerg Med. 1988; 6:439-41.

67: 73 - 9.

...................................................................................................................

Further Thoughts from the Reviewers I The authors provide a descriptive report of a medical Spanish language program for entering EM residents. While their residency cohort is small and the number of patient encounters limited, the authors’ findings are disquieting. That a significant number of errors occurred in these encounters and that the residents con- tinued to conduct clinical evaluations without the ben- efit of an interpreter in half the cases are troublesome. Also, the perils of depending upon family member in- terpretative skills alluded to in this study have been noted by others. * The ideal medical interpreter honors patient privacy, is linguistically competent in both languages, is culturally sensitive, and is medically sophis- ticated. It is unlikely that many family members will be optimal interpreters given these criteria.

Certainly at some point, the clinician who studies and practices a second language can become competent to practice in that language without an interpreter. We have yet to identify that point. Further, courses to teach Spanish as a second language to EM residents will only partially address the language needs of the emergency

Documents

Teaching Spanish to Emergency Medicine Residents