Most vendor checklists for interpreter services ask about language coverage and cost per minute. Those matter, but they are not the questions that come up during a chart review, a patient complaint, or a civil rights investigation. The questions that do are narrower: who interpreted this conversation, what credential did that person hold, was the interpreter a certified medical interpreter, and can you show it?
The patient volume is not marginal
About 26 million people in the U.S. have limited English proficiency, roughly 8% of people ages five and older. For a hospital or multi-site clinic group, that is a routine part of the daily census, not an edge case. The clinical cost of getting it wrong is measurable. A study of 1.3 million patients across 64 New Jersey emergency departments found that patients with LEP had 12% higher odds of returning to the ED within 72 hours.
What the error data says about credentials
The most cited study on interpreter accuracy in the U.S. is still Flores et al. (2012), which reviewed 57 pediatric ED encounters. Of 1,884 interpreter errors, 18% had potential clinical consequences, and the share was lower with professional interpreters (12%) than with ad hoc interpreters (22%) or no interpreter (20%). The result buyers should notice is the second one: professional interpreters with 100 or more hours of training had 2% errors of potential consequence, versus 12% for those with less training.
Two caveats. The sample is small and comes from two Massachusetts pediatric EDs, so treat the figures as directional. Also, “professional” in that study meant trained, not necessarily certified. But the pattern holds across the earlier work: an earlier pediatric study found that errors by ad hoc interpreters were more likely to carry clinical consequences than errors by hospital interpreters (77% versus 53%). Training hours and verified skill drive the outcome, and a certification is the only portable proof of either.
“Certified” now needs a follow-up question
Two bodies issue the national medical interpreter credentials: CCHI and NBCMI. Vendor language on this has gotten sloppy, and one recent change makes it worth asking directly. CCHI began sunsetting its standalone CoreCHI credential on January 1, 2025, and holders must pass a performance exam to earn CoreCHI-P and stay certified. CCHI will stop awarding CoreCHI entirely as of January 1, 2027. A vendor that describes its roster as “CoreCHI certified” in 2026 may be describing a credential that is being phased out.
Both bodies set a floor you can rely on. Each requires at least 40 hours of formal medical interpreter training. And you do not have to take the vendor’s word for any of it: CCHI runs a public registry where employers can verify certification status by name and search for certified interpreters by city, state, or language.
The legal floor is moving, so build to the stricter standard
The 2024 Section 1557 final rule set specific expectations. Its language access provisions cover qualified interpreters and translators, limits on relying on unqualified adults or children, and a Notice of Availability in English and at least the 15 most common languages among LEP residents in the relevant state. Larger providers with 15 or more staff must also designate a Section 1557 coordinator and maintain written language access procedures.
That rule has not had a quiet life. HHS’s own sample procedures document carries a note that various district court decisions have affected the 2024 rule. And two executive orders in 2025 changed the federal approach to language access, though Title VI, Section 1557, the ADA, and Section 504 remain in force. Have counsel confirm which provisions apply to your organization today. In practice, a program designed to the 2024 standard will hold up under any version of the rule, while one designed to the minimum can be caught out by the next change.
A five-point check before you sign
- Credentials, by name and language. Ask for the specific credential (CHI, CoreCHI-P, CMI, or Hub-CMI) for each interpreter and spot-check a sample in the registry.
- Specialty matching and briefing. An interpreter in an oncology consent conference should already know terms like neutropenia. Ask whether the vendor matches by specialty and whether your team can send case context beforehand.
- A signed BAA. The interpreter sees protected health information, so the vendor should be a business associate in writing, with NDAs at the individual level.
- The right modality for each setting. On-site suits scheduled consents and family conferences. Video suits urgent and telehealth encounters. Phone suits quick questions and after-hours access. Ask for average connection times by language, including lower-volume ones, around the clock.
- Documentation you can produce. Your record should show that a qualified interpreter was used, who provided the service, and how. If the vendor cannot give you usage reporting, you cannot demonstrate compliance later.
Where to look
If you are building a vendor scorecard, it helps to compare against a provider that publishes its standards. Columbus Lang, for example, describes its credentialing approach, specialty matching, and BAA practice on its medical interpreter services page, which gives you a reference point for what to ask any vendor.