About half (50%) of US adults with limited English proficiency (LEP) say they hit at least one language barrier in a health care setting over three years, and that population is not small: 28.5 million people ages five and older, or 9% of the US population, had LEP in 2024 (KFF, Overview of Health Coverage and Care for Individuals with LEP 2026). Inside hospitals the gap is measurable: in one academic surgical cohort only 12.0% of LEP patients got discharge forms in their own language (JAMA Network Open 2024), and adjusted 7-day readmission odds were 33% higher for LEP patients in a 424,745-patient New Jersey analysis (Research in Nursing and Health 2025). We aggregated data from KFF, the US Census Bureau, the Bureau of Labor Statistics, HHS rules in the Code of Federal Regulations, CCHI, Nimdzi, SEC filings, and peer-reviewed studies indexed in Europe PMC and PubMed. For the deaf and hard-of-hearing side of hospital communication, see our telecommunications relay and VRS statistics.
TL;DR
- 28.5 million people in the US (9% of those 5+) had limited English proficiency in 2024 (KFF analysis of 2024 ACS)
- 50% of LEP adults hit at least one language barrier in health care in three years (KFF Survey on Racism, Discrimination, and Health 2023)
- LEP individuals are uninsured at 23% vs 7% for English-proficient people (KFF analysis of 2024 ACS)
- Only 3.4% of LEP surgical patients had interpreter use documented at discharge (JAMA Network Open 2024)
- LEP inpatients had 33% higher adjusted odds of 7-day readmission (Research in Nursing and Health 2025)
- LEP surgical patients had longer stays in 6 of 14 studies measuring it (JAMA Network Open 2023 systematic review)
- Section 1557 requires a language-assistance notice in at least 15 languages (45 CFR 92.11)
- Executive Order 14224 (March 1, 2025) revoked Executive Order 13166 on LEP access (Federal Register 2025)
- ChatGPT produced potentially significant errors in 33.3% of Haitian Creole discharge instructions vs 8.3% for professionals (Pediatrics 2024)
- 5,504 interpreters held an active CCHI certification at the end of 2024 (CCHI 2024 Annual Report)
- Hospitals employ 7% of US interpreters and translators, at a median $63,000 (BLS Occupational Outlook Handbook 2025)
- The global interpreting market reached USD 11.7 billion in 2024; healthcare drives over 70% of US demand (Nimdzi Interpreting Index 2025)
1. Who Needs a Medical Interpreter: The LEP Population
The LEP population is growing faster than the policy attention it gets. 28.5 million people, nearly one in ten US residents ages five and older, had limited English proficiency in 2024 (KFF, Overview of Health Coverage and Care for Individuals with LEP, 2026). For hospitals, that means interpreter demand is concentrated (four states hold more than half of LEP residents) but also long-tailed, since a single children’s hospital cohort documented 95 different languages.
The language mix explains why Spanish dominates hospital interpreter staffing while remote vendors fill the tail. The Census Bureau’s most detailed language publication is older (Most recent available data: US Census Bureau, Language Use in the United States: 2019), but it shows that most people who speak another language at home are fluent in English; the clinical need is concentrated in the 38% who are not.
| Metric | Value | Source |
|---|---|---|
| People 5+ with LEP in the US (2024) | 28.5 million (9%) | KFF analysis of 2024 American Community Survey |
| Share of LEP population who are Hispanic / Asian | 62% / 21% | KFF analysis of 2024 American Community Survey |
| Primary language of LEP individuals: Spanish / Chinese / Vietnamese | 63% / 7% / 3% | KFF analysis of 2024 American Community Survey |
| LEP residents living in CA / TX / FL / NY | 23% / 13% / 11% / 9% | KFF analysis of 2024 American Community Survey |
| Medicaid and CHIP enrollees with LEP (2021) | almost 4.9 million (of 25.7 million LEP people) | MACPAC, Enrollment and Access Barriers for People with LEP, 2024 |
| People 5+ speaking a language other than English at home (2019) | 67,802,345 | US Census Bureau, Language Use in the United States: 2019 |
| Of those, share speaking English ‘very well’ / ‘well’ / ‘not well’ / ‘not at all’ | 62% / about 19% / 13% / 6% | US Census Bureau, Language Use in the United States: 2019 |
Context: the MACPAC figure (25.7 million LEP people in 2021) and the KFF figure (28.5 million in 2024) come from different ACS years, so the gap reflects growth plus sampling, not a methodology conflict. The MACPAC issue brief also found that translated Medicaid materials varied by state and document type, with paper applications more likely to be translated than online portals.
2. Language Barriers Patients Report Before and During Care
The barriers start at the front desk, not at the bedside. About half (50%) of adults with LEP encountered at least one language barrier in a health care setting in the past three years, and the most common one was paperwork (34% had trouble filling out forms), not the clinical conversation itself (KFF Survey on Racism, Discrimination, and Health 2023, reported in the KFF LEP overview).
Language-concordant care is the strongest lever in the data: LEP adults who saw a provider who spoke their language for at least half of their visits reported barriers at 40% versus 60%. Access gaps also show up before a patient ever reaches a hospital: a Health Affairs Scholar 2025 secret shopper study of Medicaid managed care practices found Mandarin speakers in Los Angeles waited 14 days longer than English speakers for a first primary care appointment.
| Metric | Value | Source |
|---|---|---|
| LEP adults reporting at least one language barrier in care (3 years) | 50% | KFF Survey on Racism, Discrimination, and Health 2023 |
| Barrier type: forms / communicating with staff / understanding instructions | 34% / 33% / 30% | KFF Survey on Racism, Discrimination, and Health 2023 |
| Barriers with language-concordant provider (half or more of visits) vs fewer | 40% vs 60% | KFF Survey on Racism, Discrimination, and Health 2023 |
| Uninsured rate, LEP vs English proficient | 23% vs 7% | KFF analysis of 2024 American Community Survey |
| Fair or poor physical health, LEP vs English proficient adults | 34% vs 19% | KFF Survey on Racism, Discrimination, and Health 2023 |
| Delayed / forgone care, LEP vs English-proficient adults (July-Dec 2020) | 49% / 41% vs 40% / 30% | Chang et al., Medical Care 2024 (NHIS, n=16,941) |
| Extra wait for first Medicaid primary care visit, Mandarin speakers in Los Angeles | 14 days | Bustamante et al., Health Affairs Scholar 2025 |
Methodology caveat: in the Medical Care 2024 study, English proficiency was not significantly associated with delayed or forgone care after adjusting for demographic, socioeconomic and health factors, so the raw gap reflects insurance and income as much as language. The same secret shopper study found Spanish and Mandarin speakers in New York waited 2.22 and 1.76 days less than English speakers, so the effect is market-specific.
3. Interpreter Use Inside Hospitals: Where the Gaps Are
The most consistent finding in recent hospital studies is not that interpreters are never used, it is that they disappear at the moments that matter most. In a 2019-2020 surgical cohort at an urban academic medical center, 88.3% of LEP patients had interpreter use documented at least once, but only 3.4% at discharge, and just 12.0% received discharge forms in their preferred language (Cevallos et al., JAMA Network Open 2024).
Clinicians themselves point to workflow, not money. In a 2024 safety-net hospital survey, cost was the lowest-ranked barrier; efficiency pressures and cumbersome access led the list (Slade et al., Joint Commission Journal on Quality and Patient Safety 2024). Structured programs do move the numbers, as a language access systems improvement (LASI) initiative and a NICU intervention bundle show below.
| Metric | Value | Source |
|---|---|---|
| LEP surgical patients with interpreter use documented at least once | 88.3% (103 of 117) | Cevallos et al., JAMA Network Open 2024 |
| Interpreter use documented within the first 24 hours | 53.0% | Cevallos et al., JAMA Network Open 2024 |
| LEP surgical patients given language-concordant discharge forms | 12.0% (14 of 117) | Cevallos et al., JAMA Network Open 2024 |
| Professional interpreter use before vs after a language access systems improvement initiative | 57% vs 66% (p=0.01) | Karliner et al., BMJ Open 2024 |
| NICU interpreter requests per Spanish-speaking patient day, before vs during intervention | 0.2 vs 1.0 (five-fold) | J Perinatol 2025 (PMID 38424233) |
| Clinician survey responses on interpreter barriers (safety-net hospital) | 221 (29.5% response rate of 750) | Slade et al., Jt Comm J Qual Patient Saf 2024 |
| Social workers vs other roles who perceive higher complication risk for culturally and linguistically diverse patients | 87.5% vs 56.8% | Slade et al., Jt Comm J Qual Patient Saf 2024 |
| Average total interpreter time per primary care encounter (in-visit plus auxiliary) | 91.1 min | Torresdey et al., J Prim Care Community Health 2024 |
Outlier note: the BMJ Open study found the lowest professional interpreter use after the intervention in visits where clinician and patient shared only partial language concordance, which is the “I speak some Spanish” problem. The full abstracts are indexed in PubMed. Clinician time pressure is a recurring theme across settings, and our emergency physician burnout statistics show why shortcuts happen.
4. Patient Safety and Clinical Outcomes
The outcome evidence is real but uneven, and honest reporting matters here. LEP inpatients had 33% higher adjusted odds of a 7-day readmission (OR 1.33, 95% CI 1.19-1.47) in a sample of 424,745 New Jersey hospitalized adults, and that gap narrowed in hospitals with better nurse work environments (Sliwinski et al., Research in Nursing and Health 2025, using 2016 data; Most recent available data for this design).
The largest surgical review found that language mostly affects the process of care (access, delays, length of stay, discharge destination) rather than mortality or complications. That pattern points to coordination failures, which interpreters and translated discharge materials can address, more than to clinical skill gaps. The JAMA Network Open 2023 systematic review covered 29 studies and 281,266 patients.
| Metric | Value | Source |
|---|---|---|
| Adjusted odds of 7-day readmission, LEP vs non-LEP inpatients | OR 1.33 (95% CI 1.19-1.47) | Sliwinski et al., Research in Nursing and Health 2025 |
| Raw 7-day readmission, LEP vs non-LEP | 4.5% vs 3.9% | Sliwinski et al., Research in Nursing and Health 2025 |
| Studies finding reduced surgical access for LEP patients | 4 of 6 | Joo et al., JAMA Network Open 2023 |
| Studies finding longer surgical length of stay for LEP patients | 6 of 14 | Joo et al., JAMA Network Open 2023 |
| Studies finding more discharge to a skilled facility for LEP patients | 3 of 4 | Joo et al., JAMA Network Open 2023 |
| Sepsis mortality, LEP vs English proficient (unadjusted) | 21.1% vs 19.6% (overall OR 1.12, 95% CI 0.88-1.42) | Limaye et al., JAMA Network Open 2024 |
| Adjusted sepsis mortality odds, LEP vs EP among non-Hispanic White patients | OR 1.56 (95% CI 1.02-2.39) | Limaye et al., JAMA Network Open 2024 |
| Pediatric postoperative mortality odds, ‘Other’ language speakers | OR 2.4 (95% CI 1.31-4.41) | Chicoine et al., J Pediatr Surg 2025 (56,655 patients) |
Context: the sepsis study found no significant overall mortality difference; the elevated risk appeared in a subgroup. In the pediatric surgery cohort (2016-2020, 95 languages), interpreter use was associated with lower odds of serious adverse events but did not reduce mortality risk. A separate 2024 quality improvement project at an academic medical center reported average monthly mortality after rapid response activation falling from 7.42% (8 of 107) to 6.09% (7 of 115) among LEP patients (J Gen Intern Med 2024, PMID 38381243), a small sample that should not be generalized.
5. The Compliance Rules and the 2025 Policy Shift
The legal floor for hospital language access got more specific in 2024 and the federal posture changed in 2025. The HHS Section 1557 final rule requires covered entities to post a notice of available language assistance in at least the 15 languages most commonly spoken by people with LEP in the relevant state (45 CFR part 92, eCFR), with a compliance deadline one year after the July 5, 2024 effective date.
Then, on March 1, 2025, Executive Order 14224 designated English as the official language and revoked Executive Order 13166, the 2000 order on access for people with LEP. The Department of Justice has since suspended operations of lep.gov pending internal review (DOJ Civil Rights Division). Title VI and Section 1557 remain law, so the hospital obligation did not disappear, but federal guidance and oversight thinned.
| Metric | Value | Source |
|---|---|---|
| Section 1557 final rule published / effective | May 6, 2024 / July 5, 2024 | Federal Register document 2024-08711 |
| Minimum languages for the notice of availability (92.11) | 15 most common LEP languages in the state | 45 CFR 92.11 |
| Categories of documents that must carry the notice | 11 (incl. consent forms, discharge papers, complaint forms) | 45 CFR 92.11 |
| Compliance deadline: Section 1557 coordinator and nondiscrimination notice | within 120 days of July 5, 2024 | 45 CFR 92.1 |
| Compliance deadline: policies, training and language notice | within one year of July 5, 2024 | 45 CFR 92.1 |
| Large-print notice font requirement | no smaller than 20-point sans serif | 45 CFR 92.11 |
| Executive Order revoking EO 13166 (LEP access) | EO 14224, signed March 1, 2025 | Federal Register 2025-03694 |
Rule details: under 45 CFR 92.201, a covered entity must not rely on a minor child to interpret except temporarily in an emergency while finding a qualified interpreter, and adult companions may interpret only in emergencies or at the patient’s documented request. Public, itemized counts of HHS Office for Civil Rights language-access complaints and resolution agreements were not retrievable for this roundup (see the Data watch note). This article is factual reporting, not legal advice.
6. Machine Translation and Remote Interpreting Quality
AI translation is moving into patient materials, and the evidence says language choice decides whether it is safe. ChatGPT produced potentially clinically significant errors in 33.3% of Haitian Creole pediatric discharge instructions, against 8.3% for professional translations (Brewster et al., Pediatrics 2024). Spanish output fared far better, which is why a blanket “AI translation works” or “AI translation fails” claim misreads the data.
The 2024 rule writes that nuance into law: when machine translation covers critical, rights-related or technical content, a qualified human translator must review it (45 CFR 92.201). For live conversation, systematic reviews keep landing on the same ranking: shared language first, then professional interpreters, with in-person or video ahead of telephone. Our speech-to-speech translation statistics cover where real-time AI translation stands outside clinical settings.
| Metric | Value | Source |
|---|---|---|
| Potentially significant errors, Haitian Creole: ChatGPT / Google Translate / professional | 33.3% / 23.3% / 8.3% | Brewster et al., Pediatrics 2024 |
| Discharge instructions tested (languages) | 20 (Spanish, Brazilian Portuguese, Haitian Creole) | Brewster et al., Pediatrics 2024 |
| Share of comparisons where professional translation was preferred: Haitian Creole / Portuguese / Spanish | 48.3% / 43.3% / 15% | Brewster et al., Pediatrics 2024 |
| Studies in a 2023 review of interpreter types and patient outcomes | 29 (in-person professional rated best) | Heath et al., J Migr Health 2023 |
| Articles in a 2024 US scoping review of LEP access and outcomes | 137 | Twersky et al., Healthcare (Basel) 2024 |
| VRI requirement under Section 1557 | real-time, full-motion video and audio over a dedicated high-speed connection | 45 CFR 92.201 |
| LEP caregivers satisfied with communication without professional interpretation (pediatric ED, Switzerland) | 44.9% | Gmünder et al., BMJ Open 2024 |
Note: the Gmünder study is Swiss and small (181 caregivers, 14 with professional interpretation), included for its direct comparison of satisfaction (5.5 vs 4.8 on its scale) rather than as a US benchmark. The Heath review concluded that any interpreter is better than none (Europe PMC record).
7. The Interpreter Workforce and Market
The workforce numbers explain why remote interpreting carries so much of the load. BLS counts 73,900 interpreter and translator jobs across all US industries in 2025, with hospitals employing only 7% of them, while one vendor alone, AMN Language Services, reported an average of 2,670 medically qualified interpreters serving more than 2,700 hospitals in Q4 2025 (BLS Occupational Outlook Handbook; AMN Healthcare 2025 Form 10-K).
Certification remains thin relative to demand. CCHI’s per-credential counts add up to 5,504 actively certified interpreters at the end of 2024 (970 + 202 + 3,877 + 222 + 233), and 3,877 of them hold the Spanish credential (CCHI 2024 Annual Report). On the market side, Nimdzi’s 2025 Interpreting Index puts global interpreting at USD 11.7 billion in 2024, with US demand primarily driven by healthcare.
| Metric | Value | Source |
|---|---|---|
| US interpreter and translator jobs (2025) / projected growth 2025-35 | 73,900 / 2% | BLS Occupational Outlook Handbook 2025 |
| Median pay, all interpreters / in hospitals (May 2025) | $60,170 / $63,000 | BLS Occupational Outlook Handbook 2025 |
| Share of interpreters and translators employed by hospitals / self-employed | 7% / 28% | BLS Occupational Outlook Handbook 2025 |
| Actively certified CCHI interpreters, end of 2024 (sum of credentials) | 5,504 (3,877 CHI-Spanish) | CCHI 2024 Annual Report |
| New CCHI certifications in 2024 / exams taken | 723 / 1,741 | CCHI 2024 Annual Report |
| Global interpreting market 2024 / 2029 projection | USD 11.7 billion / USD 17.2 billion (8% CAGR) | Nimdzi Interpreting Index 2025 |
| Healthcare share of US interpreting market | over 70% | Nimdzi Interpreting Index 2025 |
| AMN medically qualified interpreters (Q4 2025 average) / hospitals served | 2,670 / more than 2,700 | AMN Healthcare 2025 Form 10-K |
Context: CCHI 2024 pass rates ranged from 72.1% (CHI-Arabic) to 83.3% (CHI-Mandarin), with CoreCHI at 73.7%. The other national body, NBCMI, offers full certification in six languages (Spanish, Russian, Mandarin, Cantonese, Korean and Vietnamese) but does not publish a credential count on its homepage. The largest interpreting providers by 2024 revenue were LanguageLine Solutions (USD 946.0 million) and Sorenson Communications (USD 800.0 million), per Nimdzi. For the sign language side of this workforce, see our sign language interpreting statistics, and for how adults learn a second language outside clinical settings, see our language exchange apps statistics.
Summary: Hospital Medical Interpreter Services and Language Access by the Numbers
| Metric | Value | Source |
|---|---|---|
| People 5+ with LEP in the US (2024) | 28.5 million (9%) | KFF analysis of 2024 ACS |
| Spanish as primary language of LEP individuals | 63% | KFF analysis of 2024 ACS |
| Medicaid and CHIP enrollees with LEP (2021) | almost 4.9 million | MACPAC 2024 |
| LEP adults reporting a language barrier in care (3 years) | 50% | KFF Survey on Racism, Discrimination, and Health 2023 |
| Uninsured rate, LEP vs English proficient | 23% vs 7% | KFF analysis of 2024 ACS |
| Fair or poor physical health, LEP vs English proficient | 34% vs 19% | KFF Survey 2023 |
| LEP surgical patients with interpreter documented at discharge | 3.4% | Cevallos et al., JAMA Network Open 2024 |
| LEP surgical patients given discharge forms in their language | 12.0% | Cevallos et al., JAMA Network Open 2024 |
| Professional interpreter use before vs after LASI | 57% vs 66% | Karliner et al., BMJ Open 2024 |
| Adjusted 7-day readmission odds, LEP vs non-LEP | OR 1.33 | Sliwinski et al., Research in Nursing and Health 2025 |
| Studies showing longer surgical stay for LEP patients | 6 of 14 | Joo et al., JAMA Network Open 2023 |
| Pediatric postoperative mortality odds, ‘Other’ language speakers | OR 2.4 | Chicoine et al., J Pediatr Surg 2025 |
| Minimum languages for Section 1557 notice of availability | 15 | 45 CFR 92.11 |
| Section 1557 final rule effective date | July 5, 2024 | Federal Register 2024-08711 |
| EO 14224 revoking EO 13166 | March 1, 2025 | Federal Register 2025-03694 |
| ChatGPT significant-error rate, Haitian Creole discharge instructions | 33.3% (vs 8.3% professional) | Brewster et al., Pediatrics 2024 |
| Actively certified CCHI interpreters (end of 2024) | 5,504 | CCHI 2024 Annual Report |
| Hospitals’ share of US interpreter and translator employment | 7% | BLS Occupational Outlook Handbook 2025 |
| Global interpreting market (2024) | USD 11.7 billion | Nimdzi Interpreting Index 2025 |
Methodology and Sources
- KFF, Overview of Health Coverage and Care for Individuals with Limited English Proficiency (published September 21, 2026; KFF analysis of 2024 ACS and KFF Survey on Racism, Discrimination, and Health 2023)
- KFF, Designating English as the Official Language of the United States Could Impact Millions with LEP (2025, used for cross-checking)
- US Census Bureau, Language Use in the United States: 2019 (ACS-50, published 2022)
- MACPAC, Enrollment and Access Barriers for People with Limited English Proficiency (July 2024)
- US Bureau of Labor Statistics, Occupational Outlook Handbook: Interpreters and Translators (2025 data)
- Code of Federal Regulations, 45 CFR part 92 (Section 1557), via eCFR; Federal Register document 2024-08711 (API record)
- Executive Order 14224, Designating English as the Official Language of the United States (Federal Register, March 6, 2025)
- US Department of Justice Civil Rights Division, Limited English Proficiency page (lep.gov suspension notice)
- Cevallos et al., Use of Professional Interpreters for Patients With LEP Undergoing Surgery, JAMA Network Open 2024
- Joo et al., Association of Language Barriers With Perioperative and Surgical Outcomes, JAMA Network Open 2023
- Peer-reviewed abstracts retrieved through Europe PMC and PubMed: Sliwinski et al. (Res Nurs Health 2025), Bustamante et al. (Health Aff Sch 2025), Karliner et al. (BMJ Open 2024), Limaye et al. (JAMA Netw Open 2024), Chang et al. (Med Care 2024), Chicoine et al. (J Pediatr Surg 2025), Slade et al. (Jt Comm J Qual Patient Saf 2024), Torresdey et al. (J Prim Care Community Health 2024), Gmünder et al. (BMJ Open 2024), Heath et al. (J Migr Health 2023), Twersky et al. (Healthcare 2024), NICU bundle (J Perinatol 2025), rapid response QI (J Gen Intern Med 2024)
- Brewster et al., Performance of ChatGPT and Google Translate for Pediatric Discharge Instruction Translation, Pediatrics 2024
- CCHI, 2024 Annual Report; NBCMI, National Board of Certification for Medical Interpreters
- Nimdzi, Interpreting Index 2025
- AMN Healthcare, 2025 Form 10-K (filed February 20, 2026)
- Data watch: HHS Office for Civil Rights pages on Section 1557, LEP enforcement and resolution agreements returned 403 or page-not-found errors during research, and DOJ’s lep.gov is suspended, so this roundup does not report a count of language-access complaints, investigations or settlements; treat “violations” here as documented compliance gaps from peer-reviewed studies, not enforcement totals. KFF reports the share of LEP adults hitting a language barrier as 50% in its September 2026 overview and as 48% (‘around half’) in its 2025 executive order analysis, both from the same 2023 survey; we use the newer 50%. The LEP population is 28.5 million (KFF, 2024 ACS), ‘over 27 million’ (KFF, 2023 ACS) and 25.7 million (MACPAC, 2021): differences reflect survey year. Market sizing is not directly comparable: Nimdzi sizes interpreting alone at USD 11.7 billion (2024), while Slator sizes its broader language industry addressable market at USD 31.70 billion for 2025; we cite only Nimdzi for interpreting. Several studies are single-site or use older data (Sliwinski: 2016 New Jersey data; Chang: 2020 NHIS; Census: 2019), flagged as most recent available. The 2025 ACS release and the next KFF update should refresh the population figures.
Last updated: October 4, 2026. We update this roundup quarterly, and the next refresh is expected when the Census Bureau’s 2025 ACS 1-year language tables and the CCHI 2025 Annual Report are published.