Accurate Documentation as the First Step Toward Pediatric Language Justice

Effective communication with patients and their families is essential to providing high-quality health care. There is ample evidence that, when a language barrier is present, patients have decreased quality of care, including worse pain management during hospitalization, more adverse medical events, and overall worse health outcomes.1–3 This can be mitigated by access to interpreter services, which have been associated with improvement in comprehension, patient satisfaction, and clinical outcomes.4–6 From a systems level, communication via an interpreter has also been linked to decreased readmission rates and decreased hospital expenditures.7 Recent publications have also highlighted that building trust through accurate identification of patients requiring interpreter services is key for achieving language justice.8,9 However, institutions are often not successful in identifying these needs.10 The first step in ensuring access to language-equitable care is the accurate and timely identification of patients and families who require a language other than English (LOE) for communication of health care information.On a federal level, race, ethnicity, and language (REaL) data are required demographic elements that must be captured in the electronic health record.11 The Joint Commission requires the use of at least 1 question to assess English language proficiency. "How well do you speak English?" is an example of one such question that could be used. In this case, interpreter services would be offered for any response less than "very well."12,13 Additional optional questions (Table 1) are recommended to determine patients' needs for receiving either spoken or written health care information. None of these questions ascertain interpreter needs for a family unit, especially in complex cases when knowing only the patient's language is insufficient. In pediatrics, this is a scenario that is commonly encountered. Improving our ability to accurately identify both a patient's and family's communication needs in the electronic medical record (EMR) enables us to deliver equitable care that reduces the potential for error and medical harm.Recent publications have discussed special considerations for the pediatric EMR, highlighting the importance of social drivers of health in pediatric care. However, there is no specific mention of language and literacy skills as social drivers of health and the unique challenges for pediatric language documentation in the EMR.14 We know that language-concordant care improves health outcomes, and, through the 3 cases outlined, we aim to explore key considerations to undertake when delivering this care to our patients and their families. Throughout our discussions, we offer potential solutions that can be implemented at both the individual and institutional level as they apply to each case. These potential solutions are summarized in Table 2. We acknowledge that changes implemented will vary based on hospital setting, size, and resources available and hope that this discussion can serve as a catalyst for ongoing dialogue regarding potential solutions to this critical issue.Historically, the terms "preferred language" or "primary language" have been used to describe the language that a patient requires for medical discussions. Here, we will use the most inclusive term in use, ie, LOE, to describe the language needs of patients requiring interpretation or translation services.15 Additionally, to capture the diverse range of family structures, we will hereby refer to a patient's parents, guardians, and caregivers as patient families.A newborn was admitted from the delivery room for management of hypoplastic left heart syndrome. His language defaulted to "unknown" in his medical record. On review of his mothers' records, her language was documented as "Interpreter Needed: Spanish." Health care discussions occurred with interpreters for the duration of his hospitalization. How should we assign language for preverbal patients in the medical record and ensure that it is updated as language develops over time?A child with chronic kidney disease was admitted to the hospital for vomiting. His language in the medical record was listed as "English." He presented for care with his father, who spoke both English and Spanish and communicated with the care team in English. Soon after admission, the patient's father had to return to work and his mother, whose primary language was Spanish, took part in family-centered rounds. With only one language field available, there was no indication that interpreter services were needed. How can we capture interpreter needs for a verbal child with parents who speak different languages?A child was admitted with an enlarging soft-tissue mass requiring surgical excision. The patient and her mother spoke Spanish. As they were preparing the patient for discharge, the team learned that the patient's mother could not read or write. She asked that her discharge instructions be printed in Spanish so that they could be referenced by other family members involved in posthospitalization care. How can we best support families with general literacy or translation needs?In pediatrics, a broad range of patient ages lends itself to a wide spectrum of language abilities that change as the child develops. Equitable standards for language documentation in pediatrics must consider how best to document language in a patient who is pre- or nonverbal and determine how to capture changes over time. In this case, the patient's first encounter with the health care system was at birth, and his language defaulted to "unknown," without capturing his parent's language. Although there is no current universal standard, here, we discuss several possible considerations when determining best practice.In the short term, one solution is to have the language of young, preverbal children default to a primary caregiver's language. In some cases, the EMR may have the ability to link parent and child medical records. This occurs commonly for newborns, in whom a birthing parent's medical record can be easily accessed from the infant's medical record so that information such as prenatal laboratory tests and delivery data can be exchanged. In appropriate situations, language can be included in this shared information so that a child's language defaults to that of the birthing parent. Of course, this solution may not be applicable for children in the care of nonbirth parents. In this case, the family's language would have to be verified and entered in the infant's medical record manually during the first encounter with the health care system.As the child's language develops, efforts should be made to include the child in health care discussions when possible. There is no specific age when this should begin, as it is highly individual and based on multiple factors such as the child's development, social-emotional maturity, family preferences, and the diagnosis being discussed. The EMR should incorporate regular prompts for staff to confirm the accuracy of the child's documented language, at minimum with every health care encounter. Providers should consider completing additional check-ins to update and verify the documented language when patients reach appropriate developmental milestones.Adolescent patients present a special case. Federal health privacy laws state that, when minors reach age 12 years, they have a legal right to health information privacy, and personal health information cannot be shared (even with the parent) without patient permission.16,17 In some states, adolescents can provide their own consent for varying degrees of care, including testing for drugs or alcohol, screening for sexually transmitted infections, outpatient mental health care, and obstetric services. This would necessitate a physician knowing the adolescent's communication needs to facilitate health care discussions. Defaulting to a parent's language would not be appropriate and may undermine the teenager's growing autonomy. Even when a child or teenager can communicate needs on their own, it is still necessary to know the family's language. This allows questions to be addressed and consent to be obtained when necessary. For these reasons, even with adolescent patients it should be standard practice to determine interpreter needs both for the patient and family.This case discussion highlights the unique task of designating a language for a preverbal newborn and the challenges that arise as their language and communication needs change over time. Solutions implemented will vary depending on the institution. At minimum, institutions should aim to use a method of documentation that includes family language as well as regular updates of a patient's language as they age. In Case #2, we will explore potential challenges that arise when there is discrepancy in communication needs among family members.It is not uncommon to see varying communication needs within the same family, as language is highly individual. Language for discussing health care information can be discrepant at multiple levels, including between provider and child, provider and caregiver, caregiver and child, and among caregivers. The prevalence of these discrepancies is unclear given the limitations of census data-collection methods and thereby remains an area that has not been well studied.In this example, the patient's parents speak different languages. This results in different interpreter needs depending on which parent is at bedside. Here, the only language captured was that of the first family member encountered, ie, the father, with the mother's interpreter need absent. The Pediatric Health Equity Collaborative recommends that all health care systems should, at minimum, aim to document written and spoken language for the child and at least 2 caregivers for all pediatric patients.18 The Joint Commission recommends collecting the language of care for a parent, guardian, or surrogate decision-maker for any minor or incapacitated patient.14 However, they do not specify a minimum number of caregivers that should be identified or how this should be done.Again, solutions will vary based on the institution. Language is often documented in the demographics section of the medical record, but providers may not know where and how to access this buried data. Most EMRs have a top banner or side panel that is used to display key patient data so that it is quickly visible to the provider. This would be the ideal place to document interpreter needs for the patient and their family. The challenge is that there is finite space within these areas to display other important information such as allergies, code status, or weight, which may displace a caregiver's language as a priority.For the patient in our case, interpreter needs will change throughout admission depending on which family member is at bedside. Determining precise daily interpreter needs would involve frequent and repeated screening before rounds, which is not always feasible. It may be most beneficial to approach interpreter needs from a "universal precautions" perspective. In this approach, if any family member uses an LOE, then spoken and written communication should occur automatically in both languages. On rounds, providers should also make it a habit to call family members who are not able to be at bedside with the appropriate interpreter resources. This ensures that our communication optimally involves the entire family in the patient's care. It also relieves the burden of ad hoc interpretation or transmission of information by bilingual family members and aligns with federal regulations outlined by the Affordable Care Act. To implement this, providers should include interpreter needs prominently in handoff tools. Preparing for each patient encounter with the resources necessary to make discussion accessible for all family members, whether or not they are at bedside, serves to enhance patient experience and improve delivery of quality care for all patients and their families regardless of interpreter needs.4,5,8,19In Case 3, we discuss the challenges that arise when families have communication needs that extend beyond spoken language.There are vast inequities that result from unmet needs in written communication for patients and their families. In our case, the patient's mother could not read written instructions. She required that discharge instructions be completed in Spanish so that they could be referenced by other family members involved in posthospitalization care. Here, limited general literacy created a barrier to written discharge communication. The need for translation of documents into Spanish added an additional layer of complexity. This mother is not alone. One in four parents have low health literacy and rely on friends or extended family members for assistance with understanding written documents such as medication dosing or discharge instructions.20,21Operational and organizational limitations also exist in providing language-concordant discharge instructions. These include long turnaround times for translation, limited funding for translators, and poor function of translator software. These limitations are even more significant for languages without written tradition or languages of limited diffusion.21 Several studies have looked at innovative potential solutions to this problem, including pre-translated libraries of discharge instructions for common conditions, machine translation using artificial intelligence such as ChatGPT and Google Translate, or audio recordings of discharge instructions. These studies have shown promising results, but further research is needed because some of these modalities have raised concerns surrounding accuracy, reliability, and privacy.21,22As research in these areas continues to unfold, the optimal solution will likely vary depending on the resources available in each health care system. This case effectively shows the importance of avoiding assumptions regarding our patients' communication needs. Depending on the specifics of the EMR, written language would likely best be notated in the demographic section of the medical record. Ideally, this would allow translation needs to be easily visible in areas of the medical record that provide printed information for families. In some cases, this information could be used to automatically generate pre-translated instructions or, at minimum, to alert the care team when translation of documents is needed. Regardless of the solution implemented, the first step is knowledge of a patient's written language needs.Accurate documentation of a patient's and family's written and spoken language is the first step in delivering language-concordant care, especially in pediatrics. Recent publications have outlined pediatric specific considerations in the EMR but do not mention the unique challenges that language documentation represents for children and their families.11Table 2 provides solutions that may be adapted at both the individual and system levels. Pediatricians partnered with clinical informaticists are well positioned to work toward these changes. Although children are seen more frequently in their medical homes, as hospitalists, we have the advantage of a longer admission period over which to document and update language data. Hospitalization also presents an opportunity to capture these data for multiple caregivers who may be at bedside during admission but are unable to attend clinic visits because of conflicting obligations. This uniquely positions hospitalists to ensure accurate and reliable documentation of communication needs. These cases highlight key points to consider as institutions develop solutions to meet the language needs of pediatric patients and their families. Even after implementing these individual- and system-level changes, care must be taken to regularly verify accuracy of the language documented.10 Improvement in this area is a necessary building block in achieving language justice across the health care system.