Case - Bias in clinical decision making and linguistically accessible care

Bias in clinical decision making can emerge when communication feels “good enough.” In this case, a patient with limited English proficiency is assessed through partial understanding, family interpretation, and time-pressured assumptions. The case shows how language barriers, clinical shortcuts, and system design can combine to delay diagnosis — and how teams can build safer, more equitable approaches to interpreter use and patient communication.

Why this case matters

Bias is not always located in a dramatic moment of discrimination. Sometimes it appears in what seems practical: using a family member to interpret, asking yes-or-no questions, accepting nodding as understanding, or delaying interpreter use until discharge.

In clinical care, these small decisions can have significant consequences. When language access is treated as optional rather than essential, patients may be misunderstood, symptoms may be missed, and clinical reasoning may close too early.

This case invites us to ask:

  • What did the team assume they understood?
  • Where did communication become unsafe?
  • How did time pressure shape clinical judgment?
  • What system features made professional interpreter use less likely?
  • What would need to change so that language access becomes routine, not exceptional?

The goal is not to blame individual clinicians, but to examine how bias, communication practices, and system design interact.

The case

Mr. R, a 58-year-old farmworker with limited English proficiency, presents with chest discomfort and dizziness. He arrives with his adult niece who “helps translate”. The triage nurse documents “Spanish speaker, family interprets.” 

The resident obtains a history in English with frequent yes or no questions. Mr. R nods often. The ECG is non-diagnostic. He is labeled “likely GERD,” given a GI cocktail, and prepared for discharge. Before leaving, a phlebotomist who speaks Spanish casually asks if he still has “pressure.” Mr. R describes exertional heaviness, diaphoresis, and left arm pain that began yesterday, which were never elicited. Cardiac enzymes return elevated. 

The team calls language services only for discharge teaching.

What we know

  • Reliance on ad hoc interpreters increases errors, misses key symptoms, and reduces informed consent quality.
  • Clinicians overestimate their ability to “get by” in English or with partial language skills, which leads to premature closure.
  • Structural barriers include slow interpreter access, lack of phone placements, and throughput pressures.
  • Professional interpreter use improves diagnostic accuracy, patient satisfaction, and adherence, and reduces adverse events.
  • Teach-back reveals misunderstandings that simple yes or no questions miss.

How we got here

This case did not happen because one person failed to care. It happened because several ordinary practices aligned in a risky way.

  • Throughput policies valued speed over accuracy.
  • Underinvestment in interpreter infrastructure during off hours.
  • Cultural norms that treat family interpretation as “good enough.”
  • Training that focuses on vocabulary rather than communication safety.

The system made the safer option — professional interpretation early in the encounter — harder to use than the risky shortcut.

Bias Map

Bias map
LevelImplicitExplicitStructural
SelfDr. K’s fluency heuristic treats minor speech pauses as lower competenceSenior fellow’s conscious interruptions that favor familiar trainees.Rater fatigue, batch completion, unanchored scales, single-rater weight, limited direct observation.
InterpersonalLess eye contact with Amira, fewer coaching bids toward her.Differential opportunities to present or lead.Team norms that allow interruptions and sidebars.
StructuralNone of the assessment items require evidence of direct observationNarrative fields prompt personality traits.No calibration meetings, no audit of score distributions by gender, IMG status, or race.

Framework to approach the case

Use CARE-L: a five-step structure for language-accessible clinical care

CARE-L helps teams move from “we managed” to “we communicated safely.”

C: Confirm language and need early

  • At triage, document preferred language, literacy needs, and interpreter preference.
  • Trigger an automatic interpreter order in the EHR.

A: Activate the right interpreter

  • Use professional in-person or video when available, default to phone when not.
  • Never use minors. Use family only for brief comfort tasks after the clinical encounter.

R: Recenter the conversation on the patient

  • Face the patient, not the interpreter. Ask open questions first, then focused ones.
  • Use plain language and chunk and check. Avoid jargon and idioms.

E: Elicit and verify understanding

  • Use teach-back: “In your own words, what is the plan.”
  • Document the method and name or ID of the interpreter.

L: Loop back to decisions

  • Revisit differential diagnoses when new details emerge.
  • If time pressure exists, use a two-stage approach: rapid safety screen with interpreter now, full history as soon as feasible.

Practical application - Small-group activity

Ask participants to work with the case in three steps:

1. Rewrite the opening history

  • Rewrite the first five history questions for Mr. R using interpreter-ready phrasing. For example, replace: “Is it just burning after meals?” with: “Can you describe what the discomfort feels like?”; and
  • Then add a teach-back question that checks Mr. R’s understanding of the plan.

2. Build a language access bundle

Design a unit-level “language access bundle” that includes:

  • where interpreter phones or tablets are placed;
  • when interpreter use is automatically triggered;
  • how interpreter use is documented;
  • what staff should do during off-hours shortages; and
  • which metrics the unit will track.

3. Plan for real-world constraints

Draft a contingency plan for times when interpreter access is delayed. The plan should protect patient safety without normalizing unsafe shortcuts.

Debrief prompts

  • Where did premature closure happen?
  • What did the team believe they understood — and what had they not actually heard?
  • How did family interpretation shape the clinical story?
  • Which structural fixes would make professional interpreter use easier next time?
  • How should clinicians respond when family members resist professional interpreters?
  • What would it mean to treat language access as part of diagnostic safety?

Measures of success

  • Increased professional interpreter use for encounters with documented limited English proficiency.
  • Reduced diagnostic revisions after admission for chest pain among patients with limited English proficiency.
  • Higher completion rates of teach-back documentation before discharge.
  • Improved patient-reported understanding of diagnosis, treatment, and follow-up plan.
  • Improved documentation of preferred language and interpreter use at triage.