CHI common mistakes
CHI · Understand the traps and how to avoid them.
Shifting into third-person reported speech ("She says she has pain in her chest") instead of rendering the utterance in the first person the provider and patient are using.
Candidates trained in community or legal settings, or nervous about 'sounding like' the patient, default to reported speech as a safe distance; it also feels more 'professional' to some test-takers.
What to do instead
Practice with a strict first-person rule in every mock session: provider's words come out as your 'I/we', patient's words as your 'I/we'. Record yourself and flag every third-person slip; drill until the register switch is automatic.
Adding explanatory or softening words the speaker did not say, e.g. rendering 'You have cancer' as 'The doctor says there is something bad in your body, but don't worry.'
Discomfort with delivering bad news and a habit of 'helping' the patient understand lead to editorializing; candidates also fear the message will sound harsh.
What to do instead
Adopt a fidelity checklist: no additions, omissions, or hedges. Rehearse delivering difficult news verbatim in practice and debrief with a mentor on managing your own affect rather than editing the message.
Summarizing or omitting redundant-sounding segments (e.g. skipping a repeated medication list or a rambling patient history) instead of interpreting every proposition.
Time pressure and the belief that repetition is 'not important' cause compression; candidates also run out of working memory and drop content rather than ask for a pause.
What to do instead
Use note-taking for consecutive segments and practice pausing the speaker politely ('Could you pause there?'). Track completeness against a transcript in practice; treat every proposition as required.
Interpreting numbers, dosages, and units loosely (e.g. 'two times a day' for 'twice daily', dropping 'mg', or converting units without being asked).
Numbers are high-risk and easy to garble under cognitive load; candidates paraphrase units out of habit or 'correct' what they think was meant.
What to do instead
Drill number and unit accuracy separately: back-translate every dose and frequency, never convert units unless instructed, and ask the speaker to confirm ambiguous figures. Build a habit of repeating the number internally before rendering it.
Failing to manage the turn-taking: letting the provider speak for several minutes, then giving a compressed summary instead of asking for segmentation.
Deference to the provider and fear of seeming disruptive; candidates also hope they can remember it all rather than interrupt.
What to do instead
Practice scripted interventions ('Please pause after a sentence or two so I can interpret accurately'). Use them in mock encounters until they feel routine; explain the rationale to the provider at the start of the session.
Switching into advocacy or side-conversation with the patient (explaining instructions, reassuring, or answering questions directed at the provider) instead of interpreting.
Empathy and a desire to be helpful; candidates also get pulled in when the patient addresses them directly or the provider leaves the room.
What to do instead
Rehearse a boundary script: interpret the question back to the provider and redirect ('The doctor will answer that'). Debrief role-plays where the patient tries to recruit you, and practice staying in the interpreter role.
Sight translating a form line-by-line in a flat monotone without pre-reading, misreading abbreviations, or skipping headers and checkboxes.
Candidates start reading immediately under time pressure and treat sight translation as word-for-word decoding rather than a structured read-aloud task.
What to do instead
Use a fixed routine: scan the whole document first for structure, abbreviations, and numbers; then read in meaning units at a steady pace, rendering headers and checkbox labels aloud. Practice with real consent and intake forms.
Sight translating a document with dense terminology by guessing at unknown terms or leaving English words in place without flagging them.
Reluctance to admit a gap and the pressure of a continuous read-aloud task; candidates hope context will carry the meaning.
What to do instead
Build a healthcare glossary and practice the 'flag and continue' technique: state the term is unclear and continue rather than invent a meaning. Pre-read for unfamiliar terms and ask for clarification before starting when possible.
Translating a healthcare document with inconsistent terminology (same term rendered differently across the document) and without preserving formatting, headings, and placeholders.
Working sentence-by-sentence without a terminology list; candidates focus on fluency and neglect document-level consistency and layout fidelity.
What to do instead
Create and apply a termbase before translating; do a consistency pass comparing every recurring term, and preserve headings, numbering, tables, and bracketed fields exactly. Use a checklist for layout fidelity before delivery.
Delivering a translated document without proofreading or back-checking, leaving mistranslated instructions, dropped negations, or wrong dosages.
Time pressure and overconfidence after a first draft; candidates equate 'finished' with 'ready to submit'.
What to do instead
Build a mandatory QA step: read the target text alone for sense, then compare against the source line-by-line for omissions and negations, and back-translate critical instructions and numbers. Use a written QA checklist every time.