Linguistic services, including verbal interpretation and written translation, are critical to providing equitable health care for families with non-English language preference (NELP). Despite evidence of provider disuse and misuse of linguistic services and resultant adverse outcomes, few studies have assessed the practices of pediatric hospitalists related to the use of linguistic services. Our objectives were to evaluate the current practices of communication and linguistic services used by pediatric hospitalists for hospitalized children with NELP and the barriers encountered in their use.
We conducted a multicenter cross-sectional survey of pediatric hospitalist site leaders through the Pediatric Research in Inpatient Setting network, an independent, hospital-based research network. The survey was created through an iterative process and underwent a face validation process with hospitalists and a survey methodology expert.
We received responses from 72 out of 112 hospitalists (64%). Interpreter services were available widely; translation services were available in 49% of institutions. Difficulty accessing the services timely was reported as the most common barrier. Among respondents, 64% “strongly agree” or “somewhat agree” that they visit and give updates more frequently to English-speaking families than to NELP. Hospitalists reported using interpreter services “always” during 65% of admissions, 57% of discharges, and 40% on rounds.
Families with NELP do not receive appropriate linguistic care when hospitalized. Providers update English-speaking families more frequently than non-English speaking families. Future directions include optimizing workflow to reduce the time constraints on hospitalists and increasing the timeliness and quality of interpreters and translators.
Communication is critical to providing equitable health care for patients and families who prefer to communicate their healthcare information in non-English language, commonly referred as limited English proficiency(LEP) or more recently as non-English language preference.1 According to the Census Bureau data for 2018, 66 million residents in the United States speak a language other than English at home.2 The US Census Bureau defines the ability to speak English as “very well,” “well,” “not well,” and “not at all.” Anyone who reports their ability as anything else other than “very well” is considered to have limited English speaking proficiency (LEP), a terminology used by many federal agencies, including Department of Health and Human Services, Department of Justice, Department of Education and the US census Bureau.1,3 Language barriers are associated with decreased health care quality, lower health status, increased risk of serious medical events during hospitalization, including misdiagnosis, medication errors, poor patient satisfaction; and the need for repeat visits and other costly outcomes.4,5 The Joint Commission and Agency for Healthcare Research and Quality have set standards for patient-centered communication emphasizing language importance.6 In addition, multiple federal laws, including Title VI of the Civil Rights Act of 1964, the Americans with Disabilities Act, and the Affordable Care Act, require that hospitals that receive federal funds provide free-of-cost interpreter services to families who prefer to communicate in a non-English language. Despite the standards and laws, the use of interpreters remains inconsistent in hospitals.7
Current literature shows that many hospitals do not consistently provide linguistic services,7 which include interpretation services for verbal communication and translation services for written communication. Data analysis from the American Hospital Association has found a lack of uniformity in the need and offering of linguistic services, which varies by hospital location and ownership.8 National surveys and state-specific studies focusing on New Jersey and California hospitals have demonstrated insufficient access to interpreter and translation services.9–11 Pediatric providers face an additional challenge given the need to provide information in the caregiver’s language in addition to the child’s language. Two prior single-center studies conducted in an inpatient setting in children’s hospitals have reported identifying families with non-English language preference, difficult access to interpreter services, and lack of knowledge of processes and systems as critical barriers to using linguistic services.12,13 Despite evidence of provider misuse of linguistic services and the resultant adverse outcomes in hospitalized children,14,15 there are no nationally representative studies that assess the current practices of pediatric hospitalists who are directly involved with hospitalized children in regards to their practice with linguistic services.
The primary objective of this study is to evaluate the practices related to linguistic services and communication used by pediatric hospitalists for patients and families with non-English Language preference and the barriers encountered in their use. Our secondary objective was to examine mechanisms by which pediatric hospitalists identify patients and families with non-English language preference.
Methods
Data Source and Participants
We conducted a multicenter cross-sectional survey of pediatric hospitalists through the Pediatric Research in Inpatient Setting (PRIS) network, a leading pediatric hospitalist network.16 PRIS is an independent, hospital-based research network that aims to improve healthcare delivery to hospitalized children and their families and is composed of >100 sites that vary in geographic location, size, staffing models, and utilization of care practices throughout North America. The PRIS network was founded in 2010 as a hospitalist research network and has received joint support from the American Academy of Pediatrics, Academic Pediatric Association, and the Society of Hospital Medicine. The PRIS Steering Committee provided hospital-level data for all sites participating in PRIS.
The lead author’s institutional review board determined that the study did not meet the criteria for human subject research and was exempt from full review.
Survey Development and Administration
The survey included 36 questions (Appendix, survey questions), and hospitalists were asked about their individual practices and experiences at their institution related to the use of linguistic services. In addition, some questions were asked about the availability of services at their institution in general. The survey domains discussed in this article include our primary outcomes related to interpreter and translation services, our secondary objective related to the identification of language needs, and hospital and participant demographics. Survey questions on interpreter services included general practices related to interpreter use for verbal communication, verbal communication practices with patients and families with non-English language preference, and barriers. Survey questions related to translation services included practices and availability of translation services, written communication practices, and barriers. The survey included a combination of multiple-choice questions with options for free text responses. For some questions, multiple responses were allowed. We used the Likert scale and ranking questions to identify the type of interpreter and translator used (most frequent to least frequent), the barriers to their use (most common to least common), and additional ways of communication used with families with non-English language preference whose children are admitted.
We developed the survey through an iterative process, with the initial design focused on the inclusion of key topics related to the use of linguistic services identified in the literature review.11–13 Three of the questions on the survey were adapted from validated surveys published by the Agency for Healthcare Research and Quality Team STEPPS patients with LEP survey (Appendix: survey questions 9,10,17).17 The term LEP was used in the original survey distribution .Three pediatric hospitalists from different institutions who were not part of the study team reviewed the survey for face validity. These hospitalists reviewed the survey to ensure that the survey’s language, structure, and content were clear and relevant to the individual practice of pediatric hospitalists. Separately, an expert in survey methodology reviewed the content and style of the survey questions. The authors made revisions and sought feedback from the PRIS steering committee, which led to additional modifications. Finally, the survey was pilot tested with hospitalists at the lead author’s institution; additional revisions were made before the survey was distributed to the PRIS sites.
We conducted the survey between November 3, 2021 to December 10, 2021. We emailed a link with the survey attached via the Institutional Research Electronic Data Capture tool hosted at the lead author’s institution to all PRIS site leads. We requested a single response per institution and sent weekly reminders for 5 weeks.
Data Analysis
Data were analyzed using SAS 9.4 (Cary, North Carolina). Descriptive analyses (proportions and frequency) were used to capture demographic data and study outcomes. Hospital demographics were collected from the PRIS network and compared between respondent and nonrespondent sites using the χ2 test. P value of <.05 indicated statistical significance. We conducted a thematic analysis for free text answers using JMP Pro 14 (SAS Institute Inc., Cary, NC, 1989–2021). Additionally, 2 members of the study team (P.R. and M.B.) also analyzed the free text responses for descriptive analysis.
Results
Hospital and Participant Characteristics
We received responses from 72 out of 112 PRIS hospital site leads (64%). Among participants, 47% (34 of 72) identified their institution as a children’s hospital within a hospital, 96% (69 of 72) had trainees in their institutions, 68% (49 of 72) were located in an urban area, and 88% (63 of 72) reported having a dedicated language access group/division. The majority of the survey respondents were white, non-Hispanic Latino females who practiced in a general pediatric floor setting and were practicing as a hospitalist for 10 to 20 years (Table 1). When compared between respondent and nonrespondent sites, there was a difference noted between sites in terms of the type of the hospital (P = .02), the mean patient percentage of white (P = .012), and Asian (P = .014) race, and the approximate number of yearly admissions (P = .02). We did not notice other significant differences for other available characteristics (Table 2).
Demographics of Survey Participants and Their Institutions (n = 72)
Respondent Hospital Demographics . | N (%) . |
---|---|
Hospital type | |
Community hospital | 9 (13) |
Children’s hospital within a hospital | 34 (47) |
Free-standing children’s hospital | 28 (39) |
Trainees at the institution | |
Yes | 69 (96) |
No | 2 (3) |
Hospital locations | |
Urban | 49 (68) |
Suburban or rural | 22 (31) |
Dedicated language access or interpreter services group, division, or department? | |
Yes | 63 (88) |
No | 6 (7) |
Don’t know | 3 (4) |
Percentage of patients who do not speak “English very well” at local institution | |
1–5 | 2 (3) |
6–10 | 21 (29) |
11–15 | 19 (26) |
16–20 | 14 (19) |
>20 | 16 (22) |
Percentage of healthcare providers that are bilingual within division | |
0–5 | 19 (26) |
6–10 | 15 (21) |
11–15 | 10 (14) |
16–20 | 7 (10) |
>20 | 4 (6) |
Don’t know | 16 (22) |
Respondent demographics | |
Raceb | |
White or Caucasian | 49 (68) |
Black or African American | 1 (1) |
Asian | 14 (19) |
Native Hawaiian or Pacific Islander | 1 (1) |
Mixed | 2 (3) |
Ethnicityb | |
Hispanic or Latino | 1 (1) |
Non-Hispanic or Latino | 66 (92) |
Genderc | |
Male | 22 (31) |
Female | 43 (60) |
Bilingualb | |
Yes | 29 (40) |
No | 38 (53) |
Years of practice as hospitalistb | |
<2 | 3 (4) |
3–10 | 25 (35) |
10–20 | 29 (40) |
20+ | 10 (14) |
Practice settingd | |
General pediatric ward | 61 (85) |
Newborn nursery | 2 (3) |
Subspecialty inpatient ward | 2 (3) |
Others | 2 (3) |
Combination of floor, nursery, and specialty | |
Combination of behavioral health, emergency department, floor, and delivery room |
Respondent Hospital Demographics . | N (%) . |
---|---|
Hospital type | |
Community hospital | 9 (13) |
Children’s hospital within a hospital | 34 (47) |
Free-standing children’s hospital | 28 (39) |
Trainees at the institution | |
Yes | 69 (96) |
No | 2 (3) |
Hospital locations | |
Urban | 49 (68) |
Suburban or rural | 22 (31) |
Dedicated language access or interpreter services group, division, or department? | |
Yes | 63 (88) |
No | 6 (7) |
Don’t know | 3 (4) |
Percentage of patients who do not speak “English very well” at local institution | |
1–5 | 2 (3) |
6–10 | 21 (29) |
11–15 | 19 (26) |
16–20 | 14 (19) |
>20 | 16 (22) |
Percentage of healthcare providers that are bilingual within division | |
0–5 | 19 (26) |
6–10 | 15 (21) |
11–15 | 10 (14) |
16–20 | 7 (10) |
>20 | 4 (6) |
Don’t know | 16 (22) |
Respondent demographics | |
Raceb | |
White or Caucasian | 49 (68) |
Black or African American | 1 (1) |
Asian | 14 (19) |
Native Hawaiian or Pacific Islander | 1 (1) |
Mixed | 2 (3) |
Ethnicityb | |
Hispanic or Latino | 1 (1) |
Non-Hispanic or Latino | 66 (92) |
Genderc | |
Male | 22 (31) |
Female | 43 (60) |
Bilingualb | |
Yes | 29 (40) |
No | 38 (53) |
Years of practice as hospitalistb | |
<2 | 3 (4) |
3–10 | 25 (35) |
10–20 | 29 (40) |
20+ | 10 (14) |
Practice settingd | |
General pediatric ward | 61 (85) |
Newborn nursery | 2 (3) |
Subspecialty inpatient ward | 2 (3) |
Others | 2 (3) |
Combination of floor, nursery, and specialty | |
Combination of behavioral health, emergency department, floor, and delivery room |
Missing 1 datapoint.
Missing 5 datapoints.
Missing 7 datapoints.
Missing 5 datapoints.
Comparison Between Responding and Nonresponding Sites
. | R n = 72 . | NR n = 40 . | . | ||
---|---|---|---|---|---|
n . | % . | n . | % . | P . | |
Hospital classification | |||||
Community hospital | 9 | 13 | 11 | 28 | .02 |
Children’s hospital within hospital | 34 | 48 | 9 | 23 | |
Free standing children’s hospital | 28 | 39 | 20 | 50 | |
Hospital location | |||||
Urban | 49 | 69 | 33 | 83 | .12 |
Rural or suburban | 22 | 31 | 7 | 18 | |
Percentage of patient race at institutions | Mean | SD | Mean | SD | P |
White | 56 | 21 | 66 | 19 | .012 |
African American or Black | 20 | 15 | 20 | 15 | .99 |
Asian or Pacific Islander | 9 | 10 | 5 | 5 | .014 |
American Indian, Alaska Native | 2 | 6 | 4 | 7 | .12 |
Other | 16 | 17 | 14 | 13 | .39 |
Percentage of patient ethnicity at institutions | |||||
Hispanic | 22 | 17 | 27 | 22 | .25 |
Non-Hispanic | 77 | 17 | 73 | 22 | .29 |
Number of clinical FTEs comprise your pediatric hospital medicine | 16 | 14 | 16 | 17 | .80 |
Number of pediatric beds do you have in your hospital? | 199 | 166 | 148 | 115 | .06 |
Approximate number of pediatric admissions yearly | 9028 | 7937 | 6134 | 4831 | .02 |
. | R n = 72 . | NR n = 40 . | . | ||
---|---|---|---|---|---|
n . | % . | n . | % . | P . | |
Hospital classification | |||||
Community hospital | 9 | 13 | 11 | 28 | .02 |
Children’s hospital within hospital | 34 | 48 | 9 | 23 | |
Free standing children’s hospital | 28 | 39 | 20 | 50 | |
Hospital location | |||||
Urban | 49 | 69 | 33 | 83 | .12 |
Rural or suburban | 22 | 31 | 7 | 18 | |
Percentage of patient race at institutions | Mean | SD | Mean | SD | P |
White | 56 | 21 | 66 | 19 | .012 |
African American or Black | 20 | 15 | 20 | 15 | .99 |
Asian or Pacific Islander | 9 | 10 | 5 | 5 | .014 |
American Indian, Alaska Native | 2 | 6 | 4 | 7 | .12 |
Other | 16 | 17 | 14 | 13 | .39 |
Percentage of patient ethnicity at institutions | |||||
Hispanic | 22 | 17 | 27 | 22 | .25 |
Non-Hispanic | 77 | 17 | 73 | 22 | .29 |
Number of clinical FTEs comprise your pediatric hospital medicine | 16 | 14 | 16 | 17 | .80 |
Number of pediatric beds do you have in your hospital? | 199 | 166 | 148 | 115 | .06 |
Approximate number of pediatric admissions yearly | 9028 | 7937 | 6134 | 4831 | .02 |
Student’s t test for continuous data; χ2 test for categorical data. FTE, full time equivalent; NR, nonrespondents; R, respondents.
Interpreter Use and Verbal Communication Practices With Families With Non-English Language Preference
Institutions offered various types of interpreter services in the inpatient setting, including video remote interpretation (VRI) (90%), telephonic interpretation (89%), in-person interpreter employed by the hospital (78%), sign language (75%), certified bilingual staff (54%), and a certified agency contracted interpreter (42%) (Table 3). The most common type of interpreter services used by subjects was VRI (67%), telephonic interpretation (61%), followed by in-person interpreters (40%). Subjects reported the use of ad-hoc interpreters, defined as friends, family, and bilingual staff who are not certified interpreters less frequently (29% “sometimes” and 58% “rarely”). The most common barriers to using interpreter service were unavailability of interpreter services when needed (65%), time consumption while using interpreter service (61%), and not realizing before rounds that an interpreter is needed (53%) (Table 3). Hospitalists reported using interpreter services “always” during 65% of admissions, 57% of discharges, and 40% on daily rounds (Fig 1).
The use of interpreter service during admission, discharge, and rounding by pediatric hospitalists.
The use of interpreter service during admission, discharge, and rounding by pediatric hospitalists.
Availability and Use of Interpreting and Translator Services in Inpatient Setting
Availability . | N (%) . | Use of Services . | N (%) . |
---|---|---|---|
Availability of interpreting services at children’s hospitalsa | Use of interpreting service by pediatric hospitalista | Total n = 72 | |
Video remote interpretation | 65 (90) | Video remote interpretation | 48 (67) |
Telephonic interpretation | 64 (89) | Telephonic interpretation | 44 (61) |
In-person interpreter employed by hospital | 56 (78) | In-person interpreter employed by hospital | 29 (40) |
Sign language interpretation | 54 (75) | Sign language interpretation | 0 (0) |
Certified bilingual staff | 39 (54) | Certified bilingual staff | 5 (7) |
Certified agency contracted interpreter | 30 (42) | Certified agency contracted interpreter | 4 (6) |
Ad-hoc interpreter | 5 (7) | ||
Availability of preprepared documents in non-English language | Use of translation services by pediatric hospitalist | Total n = 35 | |
Patient education materials | 62 (86) | Certified hospital hired translators | 24 (69) |
Consent forms | 53 (74) | Computed assisted translation software (example SDL trados, Lokalise, Phrase etc.) | 9 (26) |
Menus for patient meals | 32 (44) | Machine translation service (example: google translate, Microsoft translation, amazon, Bing etc.) | 7 (20) |
Instructions indicating availability of language services | 26 (36) | Freelance contract translators | 4 (11) |
Don’t know | 14 (19) | ||
Other | 3 (4) | ||
Barriers to using interpreter servicesa | n (%) | Barriers to using translator servicesa | n (%) |
Interpreter service and Interpreter device not available when needed | 47 (65) | Translation service not timely | 43 (60) |
They are time consuming | 44 (61) | Poor quality of translation service | 16 (22) |
Not realizing before rounds that a family needs interpreter service | 38 (53) | Service not available | 18 (25) |
Not knowing how to access the language access service | 5 (7) | Not knowing how to access service | 20 (28) |
I do not think they contribute much | 2 (3) | Not a standard practice | 23 (32) |
Others | Others | ||
Interpreter not available in specific language | 11 (15) | Translation service not available in specific language | 3 (4) |
Family declines | 5 (7.0) | Translating all documents is not practical | 1 (1) |
Poor quality of interpreter service | 4 (5.6) | Expensive | 1 (1) |
Availability . | N (%) . | Use of Services . | N (%) . |
---|---|---|---|
Availability of interpreting services at children’s hospitalsa | Use of interpreting service by pediatric hospitalista | Total n = 72 | |
Video remote interpretation | 65 (90) | Video remote interpretation | 48 (67) |
Telephonic interpretation | 64 (89) | Telephonic interpretation | 44 (61) |
In-person interpreter employed by hospital | 56 (78) | In-person interpreter employed by hospital | 29 (40) |
Sign language interpretation | 54 (75) | Sign language interpretation | 0 (0) |
Certified bilingual staff | 39 (54) | Certified bilingual staff | 5 (7) |
Certified agency contracted interpreter | 30 (42) | Certified agency contracted interpreter | 4 (6) |
Ad-hoc interpreter | 5 (7) | ||
Availability of preprepared documents in non-English language | Use of translation services by pediatric hospitalist | Total n = 35 | |
Patient education materials | 62 (86) | Certified hospital hired translators | 24 (69) |
Consent forms | 53 (74) | Computed assisted translation software (example SDL trados, Lokalise, Phrase etc.) | 9 (26) |
Menus for patient meals | 32 (44) | Machine translation service (example: google translate, Microsoft translation, amazon, Bing etc.) | 7 (20) |
Instructions indicating availability of language services | 26 (36) | Freelance contract translators | 4 (11) |
Don’t know | 14 (19) | ||
Other | 3 (4) | ||
Barriers to using interpreter servicesa | n (%) | Barriers to using translator servicesa | n (%) |
Interpreter service and Interpreter device not available when needed | 47 (65) | Translation service not timely | 43 (60) |
They are time consuming | 44 (61) | Poor quality of translation service | 16 (22) |
Not realizing before rounds that a family needs interpreter service | 38 (53) | Service not available | 18 (25) |
Not knowing how to access the language access service | 5 (7) | Not knowing how to access service | 20 (28) |
I do not think they contribute much | 2 (3) | Not a standard practice | 23 (32) |
Others | Others | ||
Interpreter not available in specific language | 11 (15) | Translation service not available in specific language | 3 (4) |
Family declines | 5 (7.0) | Translating all documents is not practical | 1 (1) |
Poor quality of interpreter service | 4 (5.6) | Expensive | 1 (1) |
SDL, software and documentation localization.
Survey question requested to check all that apply.
Among respondents, 21% and 43% “strongly agree” and “somewhat agree” that they visit and give updates more frequently to English-speaking families than to non-English speaking families. Twenty-one percent of hospitalists reported a communication problem that created a patient safety issue during the last 6 months. Some of the patient safety examples reported by participants fell under (a) medication error, (b) prolonged length of stay, (c) missed follow-up, and/or (d) poor patient satisfaction (Table 4). Among respondents, 94% reported that they have advocated on behalf of families with non-English Language preference in the last 6 months, with 35% reporting they advocated >5 times in the past 6 months.
Examples of Communication Challenge in Patient and Families With Limited English Proficiency That Created a Patient Safety Issue
Themes . | Examples . |
---|---|
Medication error | “Home medication error or confusion and patient missed a dose of medication” |
“Incorrect antiepileptic medication ordered” | |
“Incorrect home medication reconciliation because of interpreter not being used” | |
Prolonged or unnecessary admission | “Family was not in full understanding of the diagnosis which made a long hospital stay and decisions challenging. It was a family that spoke a language that was very hard to find an interpreter for on a consistent basis” |
“A Hebrew only speaking family was ready for discharge on a Saturday- we were unable to locate any remote or in person Hebrew interpreters because of the Sabbath; we waited until the Sabbath was over and still could not identify a certified translator in Hebrew. A family member who spoke English was called to help review discharge instructions with the family once Sabbath was over. Additionally, we had to use Google translate to translate the discharge summary instructions” | |
“Patient would not be admitted if they had used an interpreter” | |
Discharge instructions | “Nearly missed important follow-up visit because didn’t understand discharge instructions” |
Poor patient care | “Consulting team did not use interpreter leading to significant misunderstanding with family” |
“Patient was unable to communicate when called nurse and so patient didn’t get desired care” | |
Poor quality of interpreter service | “Patients in COVID rooms with portable HEPA filters and no in-person interpreters for nearly 1 year (hospital policy); hospital employed certified medical interpreters used over phone but often could not hear parent or provider and vice versa” |
“The quality of phone interpreter was difficult - could not hear and kept asking to speak up” |
Themes . | Examples . |
---|---|
Medication error | “Home medication error or confusion and patient missed a dose of medication” |
“Incorrect antiepileptic medication ordered” | |
“Incorrect home medication reconciliation because of interpreter not being used” | |
Prolonged or unnecessary admission | “Family was not in full understanding of the diagnosis which made a long hospital stay and decisions challenging. It was a family that spoke a language that was very hard to find an interpreter for on a consistent basis” |
“A Hebrew only speaking family was ready for discharge on a Saturday- we were unable to locate any remote or in person Hebrew interpreters because of the Sabbath; we waited until the Sabbath was over and still could not identify a certified translator in Hebrew. A family member who spoke English was called to help review discharge instructions with the family once Sabbath was over. Additionally, we had to use Google translate to translate the discharge summary instructions” | |
“Patient would not be admitted if they had used an interpreter” | |
Discharge instructions | “Nearly missed important follow-up visit because didn’t understand discharge instructions” |
Poor patient care | “Consulting team did not use interpreter leading to significant misunderstanding with family” |
“Patient was unable to communicate when called nurse and so patient didn’t get desired care” | |
Poor quality of interpreter service | “Patients in COVID rooms with portable HEPA filters and no in-person interpreters for nearly 1 year (hospital policy); hospital employed certified medical interpreters used over phone but often could not hear parent or provider and vice versa” |
“The quality of phone interpreter was difficult - could not hear and kept asking to speak up” |
COVID, coronavirus disease 2019; HEPA, high efficiency particulate air.
Translation Services and Written Communication Practices
In our study, 36 of 66 (49%) reported that their institution provides translation services for patients and families with non-English language preference who are admitted to the hospital. The remainder reported not having translation services (24%) or not knowing if translation service was available at their institutions (19%). Of those who reported having translation services, a certified translator was used by 69% (24 of 35), followed by computed assisted translation software by 26% (9 of 35) (Table 3).
Patient education materials and consent forms in non-English language were reported as the most commonly available documents by 86% (62 of 72) and 74% (53 of 72) of subjects, respectively. Hospitalists reported that meal menus and instructions regarding the availability of language services were available 44% (32 of 72) and 36% (26 of 72) of the time, respectively. Among respondents, 8 of 67 (12%) and 25 of 67 (37%) responded that they provide patient-specific discharge instructions documents to patients and families in their preferred language, “always” and “often,” respectively. The most frequently reported barriers to providing translation services as reported by pediatric hospitalists were, “translation services not available timely” (60%), followed by “translation services not being a standard practice at their institution” (32%). Finally, the respondents reported “not knowing how to access the translation service” (28%) as the third common barrier (Table 3).
Identification of Language Needs for Patients and Families With Non-English Language Preference
Survey respondents reported various ways through which families with non-English language preference were identified at their institution. The majority reported that families are identified as having a non-English language preference when they self-identify (93%) or when informed by another team member (eg, a resident or a nurse) (92%). Seventy-eight percent reported that if during the interaction the individual does not speak and understand English well enough to participate in the conversation, they are assumed to have non-English language preference, and 76% reported having a “flag” or some form of documentation in the medical record. Among respondents, 6% reported that the documentation of a patient’s preferred language is “always” true, and 71% reported it as “often” true. When asked how the hospitalist ensured that the family’s needs for language services are met throughout the continuum of care, the most common responses were “verbally informing other team member about the family’s needs language assistance during hand-off” (64%, 46 of 72) followed by “flagging” (ie, automated banner or a flag highlighting patient’s preferred language in patient’s medical record) (54%, 39 of 72). Few respondents reported having other methods, including coordination by language access service (13%, 9 of 72), using coded bracelets, magnets, or signs on the doorway (4%, 3 of 72), and a designated place in hand-off form (6%, 4 of 72).
Discussion
Our study examined the practice of pediatric hospitalists who are on the front line of providing care for patients and families admitted to hospitals with non-English language preference. We assessed the availability of linguistic services at their respective institutions, the preferred method of using the services, and potential barriers to using them. Almost a quarter of respondents in our study reported not having a translation service in their inpatient setting, and a further 19% did not know if it was even available. In addition, only half of the participants reported that patient-specific discharge instructions (eg, discharge diagnosis, medications, follow-up) were “always or often” provided to patients and families in their preferred language.
Comparable to other studies, VRI, telephone interpretation, and in-person interpreters were the most common services available for interpretation and used by pediatric hospitalists.18 In-person and VRI are preferred by families compared with telephone interpretation and should be incorporated when possible.19,20 Interestingly, subjects reported using interpreter services “always” more frequently on admission (65%) and during discharge (60%) compared with during daily rounds (40%); this discrepancy may be because of the time-consuming nature of interpreter use reported by subjects in our study and elsewhere. Prior work has also found that interpreters are inconsistently used during rounds and that comprehension and satisfaction are higher when interpreters are used and when rounds are conducted in a language-concordant manner.21,22 Quality improvement methodologies can be used to increase the likelihood that interpreters are present for rounds.23,24 Although our study subjects reported using interpreters more consistently on admission and discharge compared with during rounds, providers should always use an interpreter. Disparities in communication with families with non-English language preference, at the time of discharge, observed in our and other studies, should be addressed to improve linguistic equity in hospitalized children.25,26
Interestingly, our study found that subjects reported being more likely to provide updates to English-speaking compared with non-English speaking families. Whereas this is not surprising given difficulties with timely access to interpreters, these finding reveal the apparent disparities among hospitalized children of families with non-English language preference. Similarly, prior work has shown that families with non-English language preference are often embarrassed or unable to directly request an interpreter, leading to poor communication and patient satisfaction20,21 Additional work is needed to determine the best mechanisms for ensuring that families with non-English language preference receive timely updates from the medical team.
Prior studies have shown that translation services are vastly underused at hospitals for patients and families with non-English language preference.27,28 Our study supported these findings. We noted that pediatric hospitalists inconsistently use written translation services or provide language-concordant written materials (eg, discharge instructions) for the patient, which is a striking disparity compared with English-speaking families. This potentially is the answer to why translation services are underused as they are either unavailable or because subjects were unaware of how to access them. Providing language-concordant written materials improves the caregiver’s understanding of the disease process and improves clinical outcomes. Potential solutions for increased provision of language concordant written materials include increased staff education on translation services, standardized instructions for common diagnoses, improving the process to request translation of materials, and use of audio-recorded instructions.27,29
Providing equitable care for families of hospitalized children with non-English language preference requires a multifaceted approach at various levels, including policymaking, institutional, and individual levels. One of the first step in ensuring that families with non-English language preference receive equitable care in the inpatient setting is correctly identifying these families. Unfortunately, the healthcare system does not appropriately identify families needing linguistic services.30 The US census question defining LEP (speak English less than very well) has high sensitivity but low specificity for predicting effective communication.31 Participants in our study noted high variability in how families with non-English language preference were identified. Most of the time, families with non-English language preference were identified when other medical team members reported needing interpreter assistance or when the family self-identified. Only a quarter of respondents reported positively that the documented language was always or often accurate. We did not ask what type of standard questions were asked by registration staff as our study focused on providers’ practice. However, the registration and intake process needs improvement in identifying families with non-English language preference to ensure that there is no delay in care or sharing of information and the time-consuming or complex operational nature of scheduling an interpreter.
Currently there are varying policies for providing linguistic care for families with Non-English Language preference based on institution, demographic factors, and state legislature. However, there is inconsistent enforcement of policies and no standards or quality metrics to accurately identify and capture language data in an inpatient setting which has led to poor documentation and reporting on the use of linguistic services. Providers, institutions, and policymakers should come together to define and standardize what constitutes an “appropriate use” of linguistic services in an inpatient setting (example: interpreter use every encounter with every physician, with every caretaker, once a day use, etc.). Similarly addressing other areas that impact the linguistic services, including reimbursement, payor policies, provider training, and impact of varying state legislature should also be researched to improve the care provided to families with non-English language preference.”32 Our study also noted that almost 20% of participants reported a patient safety problem related to communication with families with non-English language preference within the last 6 months. This is consistent with published literature showing that communication is fundamental, and families with non-English language preference are at high risk of medical errors and patient safety issues.15
Our study has limitations. First, our study is limited to site leads of the PRIS network. Though PRIS is a premier network for pediatric hospitalists, it is not a complete representation of all hospitals and pediatric hospitalists in the country. In our study, 39% of subjects identified as practicing in a free-standing children’s hospital, whereas more than 70% of all childhood hospitalizations occur at other hospital types (eg, community hospital, children’s hospital within a hospital).33 There may also be selection bias in terms of which hospitalists chose to participate in the survey; it is possible that hospitalists who are more likely to use linguistic services or have them available at their institutions were more likely to participate in the survey. Similarly, we collected individual responses from providers, and their answers may not be reflective of the institution’s practice. Since PRIS leads are academic pediatric hospitalists, we expect that they are well aware of resources within the hospitals; however, it is possible that some respondents may not have known about the linguistic resources within their institutions. The availability, use of linguistic services, and barriers were based on respondent’s reported personal practice and experiences rather than a measured outcome, which may be subject to social desirability bias. Despite most of our survey respondents being from children’s hospitals and academic centers with a dedicated language access division and having practiced for 10+years, the services provided to families with non-English language preference were not optimal. We expect the resources, awareness of resources and availability of linguistic services to be even less in a more nationally representative sample, hence, the need to address the disparities even more. Similarly, though we had 72 respondents, only 67 of them fully answered all questions. Our study also did not survey other front-line staff caring for hospitalized children, including nurses, residents, registration staff, or language access staff, including interpreters and translators, who may have differing knowledge and perception on this subject. We also noted that 40% of our respondents were bilingual, which may not represent the general population of pediatric hospitalists.
Conclusions
Families and children with non-English language preference do not receive appropriate linguistic care when hospitalized. Providers update English-speaking families more frequently than non-English speaking families. Interpreter services are available widely for hospitalists; however, their use is inconsistent, and written translation services’ availability and use are poor. Poor identification of families with non-English language preference, difficulty accessing services in a timely manner, and poor quality of services were noted as the most impactful barriers to providing appropriate linguistic services in the inpatient setting. A multifaceted approach targeting advocacy, quality improvement, and research should be taken next to make changes at a systematic, institutional, and individual level. These approaches should focus on standardizing the identification of families with non-English language preference, benchmarking linguistic services use for hospitalized children, optimizing inpatient workflow to reduce the time constraints, and increasing the timeliness and quality of interpreting and translation services in various languages.
Acknowledgments
We thank Drs Alisa Khan, Oloruntosin Adeyanju, Cory Henson, and Jane Im; and the PRIS Steering Committee for their valuable input.
FUNDING: Dr Rajbhandari received funding from a foundation grant from Akron Children’s Hospital to support PRIS survey fees.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no conflicts of interest to disclose.
Dr Rajbhandari conceptualized and designed the study, assisted with acquisition of data, analyzed and interpreted the data, and wrote the initial draft of manuscript; Drs Glick and VanGeest conceptualized and designed the study, and analyzed and interpreted the data; Ms Brown assisted with acquisition of data and analyzed and interpreted the data; and all authors reviewed and revised the manuscript, approved the final manuscript as submitted, and agree to be accountable for all aspects of the work.
COMPANION PAPER: A companion to this article can be found at www.hosppeds.org/cgi/doi/10.1542/hpeds.2022-007077.
DATA SHARING STATEMENT: The deidentified data that support the findings of this study are available on request from the corresponding author.
Comments