Outsourced To Kids: Childhood Language Brokering in Texas

When Onelia Navarro underwent serious surgery, she was the one to tell her parents that she might not survive the procedure—not the doctors. The hospital did not provide translators for her Spanish-speaking parents. She was 11 years old.

“The hospital at the time did not offer interpreting services to us because they thought I could handle the interpreting portion,” Navarro recounted in an essay about child interpreters. “That day, I had to inform my parents that there was a high chance of my not surviving this procedure, as the survival rate in my condition was very low. My parents did not get the space they needed to grieve, considering what they were about to experience with their child. They also avoided asking scary questions they probably needed answers to as a way of protecting me. I remember that my biggest fear was that if I did die during the procedure, my parents would be isolated from the outside world because there was no one else to interpret for them.”

Navarro’s experience is far from isolated. According to Children at Risk, a Houston-based advocacy group, 1 in 3 Texas children belongs to an immigrant family. Texas is home to 14 percent of the total population designated “LEP,” or “Limited English Proficiency.” Federal laws mandate that LEP individuals are entitled to competent translation services in medical, legal, and educational settings. Yet these protections only extend as far as budgeting and staffing limits allow, leaving a distinct gap between paper and process. In reality, children are stepping in as de facto interpreters—a phenomenon known as childhood language brokering.

A collaborative study between Texas State University, UT Austin, and the Children’s Hospital of San Antonio cites that “more than 70 percent of adolescents from Mexican immigrant families act as language brokers for their families.” An earlier study indicates that about 70–80 percent of Asian American adolescents act as language brokers for their parents. Children in immigrant families often begin language brokering in early elementary school, just after they’ve begun to grasp the fundamentals of the English language—and well before they’re developmentally prepared to interpret complex or emotionally charged information.

“I was usually praised for being a 'mature and helpful' child,” Navarro writes. “However, people did not realize how stressful this was for me. I was…an 11-year-old kid, and I was constantly struggling due to uncertainty about my ability to deliver messages accurately in both languages. I had to grow up quickly to be able to understand adult life without getting the chance to process the difficult information I was being asked to interpret.”

Taken together, the available data suggests that there are as many as 1.5 to 2 million child language brokers in Texas. Yet, no official statewide count exists.

The Law Ends at the State

Dr. Alice Chen, a physician and public health advocate, published an article detailing the story of Gricelda Zamora, a 13-year-old girl:

“She served as her family’s interpreter. When she developed severe abdominal pain, her parents took her to the hospital. Unfortunately, Gricelda was too sick to interpret for herself, and the hospital did not provide an interpreter. After a night of observation, her Spanish-speaking parents were told, without the aid of an interpreter, to bring her back immediately if her symptoms worsened, and otherwise to follow up with a doctor in three days. However, what her parents understood from the conversation was that they should wait three days to see the doctor. After two days, with Gricelda’s condition deteriorating, they felt they could no longer wait, and rushed her back to the emergency department. Doctors discovered she had a ruptured appendix. She was airlifted to a nearby medical center in Phoenix, where she died a few hours later.”

A report from the National Health Law Program (NHLP) found that failures to provide qualified medical interpreters have been linked to serious medical errors, including patient death. Zamora’s story is closely echoed in the Tran family’s case.

“The patient, a 9-year old Vietnamese girl, died from a reaction to the drug Reglan. Her parents primarily spoke Vietnamese, yet no competent interpreter was used throughout Ms. Tran’s encounters with the medical system. Instead, records show that the 9-year-old patient and her 16-year-old brother served as interpreters during the medical encounter…The case demonstrates the failure to provide language access on multiple fronts—the failure to utilize a competent interpreter, the use of a minor child as an interpreter, and the lack of a translated informed consent form."

Access to competent translation in medical care is one of the U.S.’s most longstanding protections, codified in federal law more than half a century ago. Under Title VI of the Civil Rights Act, any hospital, school, or agency receiving federal funding is prohibited from discriminating “on the grounds of national origin”—a category that federal courts and agencies have long interpreted to include language. Dr. Chen writes that the 1964 act is “the single most important piece of legislation for providing LEP individuals a legal right to language assistance services.”

At the time, the law required “meaningful access” to information. In practice, the definition of “meaningful” varied and was enforced upon complaint, which required access to additional (often untranslated) materials. Federal guidance provided a clear definition: qualified interpreters and translated materials. Yet, without consistent oversight or funding, many institutions relied on the most readily available translators: patients’ children.

The use of untrained interpreters—particularly children—had been repeatedly implicated as a risk to patient safety and barrier to quality care, thus requiring new legislation. When the Affordable Care Act was passed in 2010, Section 1557 was authored to prohibit the practice of child language brokering in medical settings. It explicitly states that providers cannot rely on minor children to interpret except in true emergencies, and even then, only until a qualified interpreter becomes available.

The Texas Health and Human Services Commission states that it is “required to provide interpreter services and written translated materials to applicants and recipients who are Limited English Proficient (LEP),” but this requirement applies only to programs administered by the agency—such as Medicaid and other public benefits—not to the broader network of private hospitals, clinics, and providers operating across the state.

At least 43 states have adopted some form of unified language access law or policy to be applied across both public and private healthcare settings. Texas—where 14% of the nation’s LEP population resides—has not.

The Limits of “Meaningful Access”

The law is clear: children are not to be used as ad-hoc translators. Yet the NHLP reports that the practice remains common and is even less commonly reprimanded.

Part of the problem is that “meaningful access” is enforced through a flexible standard rather than a uniform system, leaving institutions to interpret their obligations differently in clinics than in courtrooms. For some Texas hospitals, this means translators with comprehensive education and extended knowledge of medical terms. Others post job openings for translators, regardless of qualifications. The Harris County Medical Society advises its physicians that “a certified interpreter does not need to be provided and is not required by the ADA - it requires only a qualified interpreter. This is potentially a cost-saving distinction.”

Hospitals, facing cost and staffing pressures, have increasingly turned to tele-interpretation services rather than hiring on-site human translators. Houston Methodist bills this as “innovative new technology,” describing how over-the-phone interpretation is “the most widely used option at Houston Methodist, connecting staff to interpreters across more than 200 languages. In practice, nurses and physicians simply place the call, select the language, and put the interpreter on speaker for real-time participation.” However, patient experience defies the sunny claims of marketing materials.

“When I asked if I could be with her in the room to translate, she said no because it's a small room and they have a tele-translator on standby,” one Reddit user reported. “I told her that we would really feel much better if I could translate only because we've had experiences with translators before that usually they speak our language well, but they don't translate back to English properly (causing misunderstandings and information to be mistranslated). The nurse said no and told me to wait and I said okay. She came back and said, actually, you can come because the translator isn't answering."

Though “meaningful access” to language services is documented in only 13% of U.S. hospitals, enforcement of the federal law is largely reactive: violations are often addressed only after a patient or family files a complaint with the federal Office for Civil Rights. As a result, language access violations are rarely documented and addressed. Filing a complaint requires a patient or family to recognize that something went wrong, understand that it may have been illegal, and then navigate a federal reporting process that is often conducted in English or requires documentation that has not been translated, either.

Could AI Be The Answer?

As the number of LEP patients increases with Texas's growing immigrant population, the pressure to find cheaper solutions has only intensified; demand is expected to grow 20 percent by 2031. Texas has no Medicaid reimbursement for interpreter services and no statewide certification standard, which means hospitals here have less financial incentive to staff for language access than in nearly any other state in the country.

Interpreters are expensive—tele-interpretation runs up to $3 a minute, video interpretation $3.49, and in-person interpreters between $25 and $150 an hour. With hospitals understaffed and underfunded, administrators have begun exploring AI alternatives. Section 1557 requires competent translation, meaning that AI can assist, but it does not meet the standard of care on its own. It also struggles with emotional nuance, medical terminology, and—vitally—clinical context. Additionally, the tools perform unevenly across languages. One study explicitly states that medical discharge instructions translated into Spanish are more accurate than those translated into Mandarin Chinese.   

This creates a landscape in which new technologies have created a new burden: when the tele-translator doesn’t pick up, when the translation is inaccurate, or the machine is less than emotionally nuanced, someone must still step in to close the gap. And in many cases, that person is not a trained interpreter, but the bilingual family member sitting in the room.

Often, the child.

Not All Child Language Brokers Are the Same

Many firsthand accounts posted online detail negative experiences with childhood language brokering:

“At the age of 9, I was expected to understand what housing, bills, utilities, etc., statements all mean and translate it to my parents so that they can make the most informed decision. Of course I never understood anything the forms said so I would confess, telling them that I don't understand. They would then get angry and belittle me, saying things like "I send you to school and you can't even read a damn paper?"

“I'm currently sitting in a waiting room with my parent because he can't articulate or understand the doctor so I have to sit there and translate. Iiterally do everything too. I'm a child of immigrants and I wish that my parents would have taken the initiative to go to school to learn the basics and navigate life here instead of piling it all on the kids.”

“Imagine taking your parents to a doctors office and they try to explain what's happening. I don't even understand what they're saying and my dad is rushing me to translate…I can't speak my native tongue very well so most of it is broken. I also have a hard time communicating with my parents about common things and they somehow expect me to translate medical terms?”

Yet, childhood language brokering does not universally damage children. For twelve years, Dr. Su Yeong Kim watched children carry this weight and tried to understand what made some of them stronger for it. Kim, a professor of human development at the University of Texas at Austin, began her landmark study on childhood language brokers in 2012. She surveyed hundreds of central Texas adolescents of Mexican origin—kids who were already navigating the space between two languages. Kim followed them through middle school and into adulthood, tracking not just what they translated but how they felt about it: whether they felt competent or overwhelmed, proud or ashamed, seen or abandoned to the task.

What she found complicated the narrative. Language brokering can be harmful—but it doesn’t have to be. When a child is praised for their contributions, the experience of translating for their parents can enhance family bonding, self-esteem, and bilingual fluency. The children who fared best were those who had both the language skills and the support systems to navigate effectively. Positive experiences like these are rarely achieved in high-stress scenarios such as medical crises.

The children who fared worst were the ones asked to translate under extreme stress (like hospitalizations) and felt that they didn’t have the words or emotional skillset to navigate their parents’ responses.

Kim's after-school program at Ojeda Middle School in Del Valle ISD was built on that insight. Project SEED is the only community-based intervention of its kind in Texas. It was designed not to eliminate language brokering, which for many families is impossible, but to give children and families the tools that transform the experience from exploitation into universal gain.

Kim’s own experience with childhood language brokering was positive. In a podcast interview conducted within UT’s College of Natural Sciences, Kim said that “for me, it really helped me have a stronger bond with my parents, because I was helping them, and because they were always relying on me, I felt really important for my family and I felt like everything was really positive and thought it was so important to do.” However, she goes on to elaborate that while she “derived so many positive benefits from doing it for my family, now I know that's not the typical story—because that's not what the empirical data shows—there's a big range of how adolescents feel about language brokering from the positive to the negative.”

A Disappearing Body of Research

Kim’s research is exceptionally rare. The empirical record she built over twelve years building hosted its final data collection in December 2024. The study had shrunk from 604 to 280 families, and funding had shrunk even further. In the Jun. 2025 interview, Kim stated that it was “unlikely” that she could secure another grant, citing “recent federal funding cuts.” The single most important long-term dataset on childhood language brokering in Texas is over—due to budget cuts.

The NIH, NSF, the Russell Sage Foundation, the Spencer Foundation, and the Hogg Foundation for Mental Health funded Kim’s research. The National Institute on Minority Health and Health Disparities lost nearly 30% of its NIH grants between Feb. and Apr. 2025 alone. As of November 2025, the NIH and the NSF had collectively frozen or terminated more than 3,800 grants—roughly $3 billion in unspent research funds. Research on ethnic minority youth, immigrant families, bilingualism, and cultural stressors—all data fields used to shape policy—have been flagged and, therefore, terminated.

In Feb. 2026, UT Austin announced that it would be consolidating seven ethnic and gender studies departments into only two. The University of North Texas is “phasing out” its entire linguistics department and eliminating more than 70 programs, many of which are related to DEI initiatives. The gap in the academic record is being manufactured by frozen federal funding. In the same moment, the conditions that produce child language brokers—mass enforcement, chilling effects on healthcare attendance, AI tools being deployed as cheap substitutes—are increasingly present.

Child Language Brokers Shoulder the Cost

Section 1557 was written because the medical evidence was overwhelming: children should not be forced to translate in stressful situations. Fifteen years later, they are still in Texas waiting rooms with their parents. Onelia Navarro survived her surgery. The law that should have protected her was already in the books when she was 11, in that room, doing the math on her own survival odds. It was there when Gricelda Zamora was not so lucky. It is there now, while Texas's research infrastructure dissolves and its hospitals shop for cheaper apps.