A Vietnamese-speaking grandmother gets the after-visit summary in English. Her daughter, who took her to the cardiology appointment and speaks both languages, becomes the translator. By the time the daughter has translated the medication change, the follow-up date, and the open question to ask next time, the grandmother has been managing her own care from a 4-page printout of jargon she can't read for about 3 hours. The federal regulation that requires the hospital to provide an interpreter during the visit ran out the moment the visit ended.
This is the bilingual-care gap. The law gets you to the door of the exam room. It does not follow you home.
The legal floor (which is real, and often misunderstood)
Three pieces of federal law set the floor for language access in US healthcare:
Title VI of the Civil Rights Act of 1964 prohibits discrimination on the basis of national origin in any program receiving federal financial assistance — which means essentially every hospital, every clinic that takes Medicare or Medicaid, every Federally Qualified Health Center. Title VI has been interpreted, since the LAU v. Nichols Supreme Court decision in 1974, to require recipients of federal funds to take reasonable steps to provide meaningful access for people with limited English proficiency (LEP).
Section 1557 of the Affordable Care Act extends and operationalizes this. Specifically, it requires covered entities (most healthcare providers and any health-insurance program receiving federal funds) to:
- Provide qualified interpreters at no cost to LEP patients
- Provide written notices and translated taglines in the 15 most-common non-English languages spoken in the state
- Post non-discrimination notices and taglines in 15 languages on websites and in physical locations
- Not require LEP patients to bring their own interpreter, and not rely on minor children to interpret
Executive Order 13166, signed in 2000, directs federal agencies to develop language-access plans. Most importantly, it required HHS to publish LEP guidance that healthcare entities are required to follow.
This is a substantial framework. It is genuinely better than the framework in most of the world. A Spanish-dominant patient walking into a US hospital in 2026 is entitled to — and in most circumstances actually receives — a phone interpreter, an in-person interpreter for complex visits, translated discharge instructions, and translated consent forms. A bilingual sign at the registration desk lets them know how to ask. Major health systems have invested in language-access programs because they have to and because patient safety outcomes improve when they do.
But notice what the framework covers, and what it does not.
The gap, in three layers
Layer 1: The clinical encounter is covered. Once you are in the room with the clinician, you have a right to an interpreter. The clinician is talking to you in your language (through the interpreter). The clinician is documenting your visit in their (English) electronic health record. The legal obligation is satisfied at the moment of the encounter.
Layer 2: The post-visit handoff is partially covered. Written discharge instructions are required to be in your language (or you must be offered a translation). After-visit summaries are increasingly available in patient-portal-translated form. Section 1557's 15-language requirement gives most major-language LEP patients access to translated standard documents.
But the after-visit summary is the snapshot from one visit. The medication list across all your specialists, the lab trends over the past year, the prior-authorization correspondence with the insurance carrier, the patient-portal message from your nephrologist asking about your weight, the appointment-reminder text from the cardiology clinic — these arrive in whatever language the system defaults to, which is overwhelmingly English. Even in major Spanish-speaking metros, most patient portals offer Spanish interface translation but not Spanish content translation. Mandarin, Vietnamese, Korean, Tagalog, Russian, Arabic, Polish, Haitian Creole — the deeper the languages list goes, the thinner the coverage.
Layer 3: The family's caregiving coordination is not covered at all. This is where most bilingual American households actually live. The daughter who translates everything for her mother. The son who texts the family chat in English and his parents in Mandarin. The grandson who interprets the medication change to his grandfather over dinner because the discharge instructions came in English and the formal interpretation services ended when the hospital visit ended.
The bilingual caregiver — the daughter, the son, the grandson — is doing translation work that has no legal protection because there is no provider-side encounter to attach the legal protection to. The work is real, invisible, ongoing, and falls heaviest on the second-generation American who grew up bilingual and inherited the family's English-speaking-world duties. The hospital fulfilled its legal obligation; the family is still carrying the load.
What this looks like in actual households
A few patterns we see, reading the family-caregiving literature and listening to families directly:
The eldest-daughter translator phenomenon. In many immigrant families, one family member ends up being the de-facto translator for an aging parent's care. Usually it's the eldest daughter; usually it's structural rather than chosen; almost always it's on top of her existing job and her own family's caregiving. The cumulative time spent on translation — appointments, after-visit summaries, medication explanations, insurance correspondence, pharmacy interactions — runs into hundreds of hours per year for active caregiving relationships. Almost none of it is compensated; almost none of it is visible to anyone outside the family.
The minor child interpreter (which is technically banned). Section 1557 explicitly prohibits requiring LEP patients to bring their own interpreter, and forbids using minor children as interpreters. In practice, this happens constantly — a 12-year-old interpreting for a grandmother at a pediatrician visit, a 16-year-old explaining a Medicare denial letter at the kitchen table. The rule is not enforced by anyone with skin in the game outside the encounter itself, and outside the encounter the prohibition does not apply.
The English-portal lockout. A Spanish-dominant patient with proxy access set up by their bilingual daughter logs into their patient portal. The interface translates to Spanish. The content of every lab result, every doctor's note, every appointment summary stays in English. The patient can navigate the portal but cannot read the chart. The daughter ends up screen-sharing every visit. The system technically complies with language access; the patient is functionally locked out.
The "use Google Translate" suggestion. Well-meaning clinicians sometimes suggest patients and families use Google Translate for the materials that aren't natively translated. Google Translate has improved dramatically — but medical contexts are exactly where machine translation accuracy still has its largest gaps. Misinterpreting "take with food" vs. "do not take with food" is the kind of error that affects safety. Recommending Google Translate as a substitute for professional translation in a clinical context is a known shortcut and a known risk.
What the depression conversation looks like in a bilingual household
We added a Caring for someone with major depression page this week. The Stanley-Brown Safety Plan we reference is one of the most-evidence-backed interventions for suicide risk in a mental health context. The plan is in English. Translated versions exist in some languages, by some research groups, in some PDFs, but the patient is rarely handed a translated version at the clinical encounter that produced the plan. The family member who has to translate the safety plan for an LEP family member at risk is the same family member who is most worried about the situation. The translation work and the emotional work compound.
A similar pattern shows up in the Medicaid work-requirement implementation we wrote about this week: Section 1557 requires the outreach materials to be translated into the languages of the populations served, but enforcement is uneven across states, and many state Medicaid agencies are publishing the first round of outreach materials in English + Spanish only. The Mandarin-speaking grandmother whose adult daughter is her family caregiver gets the letter in English. Either the daughter translates it, or the family misses the exemption deadline, or both.
What families can do, today, with what exists
There are real leverage points, even within the gap. A few patterns that work:
Use the right to a phone interpreter, every time, including for non-emergency calls. Title VI applies to phone calls placed by the patient or family to the provider's office, not just to in-person visits. You can call the cardiology office and ask for a phone interpreter. The office is supposed to provide one. Many family caregivers don't know this, and the office staff sometimes don't either; insist when you have to.
Ask for the after-visit summary in the patient's preferred language, in writing. Even when the verbal interpreter was provided, the written summary often defaults to English. Section 1557 standard-document translation requirements mean you can ask for the visit summary in the patient's language, and the office is supposed to provide it. Push for it.
Document the language gap in the patient's record. Most EHR systems have a "preferred language" field, but it's often left blank or set to a default. Ask the registration desk to update it. Once the field is correctly set, downstream systems (appointment reminders, portal interface, lab letters, refill notices) follow it.
Translate at the family-coordination layer, not just the clinical layer. The legal framework will not extend into your family's daily caregiving work; that's the gap. Tools that translate the family workspace itself — the visit summary you write up after the appointment, the medication change you log, the question you want to ask next time — make the family-coordination work bilingual at the layer where families actually live. (Kintaria does this in 7 languages, side-by-side; the English version stays the canonical record, the parent reads their language at the same time. Other tools do versions of this too. The category is small but the families it serves are not.)
File a complaint when the legal floor isn't met. HHS Office for Civil Rights takes complaints about Section 1557 violations. The complaint process is online, free, and does not require an attorney. The deadline is generally 180 days from the incident. Cases that produce systemic change — a major hospital changing its translated-materials practices — usually start as one family's complaint.
What policy could do better
A short list of policy moves that would close meaningful pieces of this gap:
Extend Section 1557 written-document requirements deeper into post-visit + portal content. The 15-language framework for standard documents is real. Extending it explicitly to portal chart-note content, AVS letters, lab-result narratives, and appointment-reminder text would address the patient-portal lockout pattern.
Fund family-caregiver translation support directly. The work that bilingual family caregivers are doing has clear economic value (avoiding readmissions, improving medication adherence, reducing the cost of the otherwise-required clinical interpreter calls). Federal or state caregiver-stipend frameworks could explicitly include translation labor.
Strengthen the "no minor children as interpreters" enforcement. The Section 1557 prohibition is real but rarely enforced outside the formal clinical encounter. Discharge planning, registration, and front-desk interactions are exactly the moments where the prohibition gets violated in practice.
Pilot AI-translation infrastructure with appropriate guardrails. Machine translation has improved enough to be useful in many caregiving contexts; the risks are well-known and addressable with the right design (medical-terminology tuning, side-by-side original-preserved presentation, human review for high-stakes contexts). A federal pilot program could fund this for community-org settings that serve LEP families directly.
The 60-year history of language-access civil rights in US healthcare is one of the genuine successes of the post-1964 civil-rights framework. The next 10 years of work is extending what was won in the clinical encounter into the family-coordination work that families actually do. Until that work is funded, designed, and built, the daughters and sons and grandchildren of LEP elders will carry the load. They have been carrying it for decades. They deserve better tools than the discharge printout and a Google Translate tab.