Medical interpretation is billed by the minute. The encounter, however, is not entirely made up of conversation.

‍

In every clinical encounter with a non-English-speaking patient, there are two distinct phases: communication time (provider and patient exchanging questions, answers and instructions) and what we call silent clinical time (documentation, chart review, ordering and physical examination). Remote interpreters, whether on video or over the phone, are typically attached to the encounter from start to finish, billing through both phases. The result is a structural inefficiency that is largely invisible in language-access program budgets but present in every single encounter. Understanding where that waste originates, how large it is and what options exist to address it is now a practical operations question, not a theoretical one.

‍

1. How Much of a Medical Encounter Does Not Require an Interpreter?

A 2020 study published in JAMA Network Open analyzed EHR usage across outpatient encounters and found that physicians spent an average of 16 minutes and 14 seconds per encounter on EHR tasks: chart review (33%), documentation (24%) and ordering (17%). That is more than 16 minutes per visit during which the physician's attention is directed at a screen, not at a patient conversation. For any encounter with a live remote interpreter, those 16 minutes are billed at the same per-minute rate as active dialogue.

‍

A separate study of outpatient encounters in Turkey, published in PubMed Central, measured total visit duration and time breakdowns across encounter steps. Out of an average encounter of 22.7 minutes, approximately 10.3 minutes did not require active interpreter communication. That is roughly 45% of the encounter.

‍

The specific numbers vary by specialty, care setting and patient acuity. In an emergency department triage encounter, silence time may be shorter. In a primary care visit involving a complex medication reconciliation or an optometry exam, it may be longer. But the structure is consistent: every medical encounter has non-verbal clinical time, and the current billing model for remote interpretation does not distinguish between active and silent minutes.

‍

2. What Providers Are Actually Doing About It

In interviews with providers across specialties, two patterns emerge.

‍

The first is manual management. An MD from an inpatient hospital in New York described the experience directly: "If I don't say anything for 15 minutes, after 15 minutes they'll hang up. I have to request it. Yes, I have to request every 15 minutes for them to stay on the line." This is not an anomaly. It is a workaround that shifts the administrative burden to the clinician, who must interrupt their EHR workflow to re-authorize interpreter presence on a timed loop.

‍

The second is selective deployment. An MD at an optometry community health center described a different approach: "We call an interpreter over the phone for the beginning part of the visit but since that service charges by the minute, we can't have the interpreter on the phone for the whole time." In their setting, the optometry exam occupies most of the visit. The interpreter is present only for the history-taking segment, then disconnected. This reduces cost but creates a gap: if a clinical finding requires real-time explanation mid-exam, the interpreter is no longer available.

‍

Neither approach solves the problem. One optimizes for cost, one optimizes for coverage and both require active clinical judgment about when to start and stop the meter. That judgment is not neutral. It introduces variation and, in some cases, risk.

‍

3. Why This Is Structurally Invisible in Program Budgets

Language-access program budgets are typically reported in aggregate minutes and total spend. A health system reviewing an annual contract sees total volume and total cost. It does not see what percentage of those billed minutes were active versus silent. That data is not surfaced by most traditional language-service providers because it is not in their interest to surface it.

‍

The result is that the true cost of interpretation per communication-minute is materially higher than the per-minute rate on the contract. If 45% of billed time is silence, then for every 100 minutes billed, 55 minutes are doing real language work. The effective cost per communication-minute is approximately 1.8x the headline rate.

‍

For a community health center running 30,000 interpreter minutes per year, that gap compounds quickly. It is not a rounding error. It is a structural overpayment that scales directly with volume.

‍

4. What a Better Architecture Looks Like

The fix is not to deploy fewer interpreters or to ask clinicians to manage interpreter access more carefully. Both of those approaches shift burden and increase risk.

‍

The more defensible path is to rethink the session model itself. An on-demand, instant-connect interpretation infrastructure, rather than a continuous attached session, allows the provider to summon interpretation at communication moments and disengage during silent clinical time without breaking the workflow or requiring active management.

‍

This is precisely where hybrid AI-plus-human interpretation programs show operational advantages. For the 2026 NACHC/ScaleHealth Accelerator cohort, No Barrier was selected as one of six companies supporting language-access infrastructure for the 52 million Americans served by Community Health Centers. Part of the selection rationale was the platform's ability to reduce per-encounter interpretation cost without reducing coverage. AI interpretation handles high-volume, short-communication interactions with sub-second response times. Human interpreters are available within the same interface for emotionally complex conversations, consent discussions and situations where dialect precision is critical.

‍

A 2026 NEJM Catalyst study evaluating AI interpretation at Brigham and Women's Hospital found that patients did not view AI and remote video interpretation as competing options. They viewed each as appropriate for different clinical contexts, with AI preferred for speed and privacy and human interpretation preferred for complex or sensitive conversations. That framing also applies to the session-structure problem: AI interpretation, by design, does not have an idle billing clock. A query-response model eliminates silence billing structurally.

‍

Diana Erani, COO of NACHC and a guest on the Care Culture Talks podcast, has described the operational reality for CHCs plainly: the organizations serving the highest proportions of LEP patients are often those with the thinnest operational margins. Every dollar recovered from structural waste in interpretation billing is a dollar that can extend clinical capacity.

‍

The Practical Takeaway

Silence time in medical encounters is not a problem that can be fully eliminated. Chart review, documentation and physical examination are clinical necessities. What can change is whether those minutes continue to be billed at per-minute interpretation rates.

‍

Health systems and community health centers reviewing their language-access program economics should ask their current vendor one specific question: what percentage of billed minutes in the past 12 months reflected active spoken interpretation versus interpreter presence during silent clinical time? If that data is unavailable, the program is managing to a rate, not to an outcome.

‍

For organizations ready to move toward a session model that separates communication time from silent clinical time, the economics are now available and the technology to support it is in production deployment across FQHCs, hospital systems and specialty practices.

‍

‍
‍

Resources:

¹ _{https://pubmed.ncbi.nlm.nih.gov/31931523/}
² _{https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5541965/}