Surgeons not fluent in non-English languages often used their limited skills to obtain preoperative informed consent from LEP patients

Description #

In this survey of surgeons at a Boston academic medical center, surgeons who were not fluent in non-English languages reported they often used those limited language skills to obtain preoperative informed consent from their LEP patients. About one-third of respondents (n = 50, 32%) reported using their own non-English language skills to consent patients for surgery, and nonfluent surgeons reported doing so even when a professional interpreter was readily available.

"Our study found that surgeons who were not fluent in non-English languages reported they often used those limited skills to obtain preoperative informed consent from their LEP patients." (Patel, 2016, p. 520)

"About one-third of respondents (n = 50, 32%) reported using their own non-English language skills to consent patients for surgery." (Patel, 2016, p. 517)

"Nonfluent surgeons consistently reported using their non-English language skills, even when a professional interpreter was readily available." (Patel, 2016, p. 518)

Methods Context #

What? #ⓘ

The observable: surgeon-reported use of their own (self-rated nonfluent) non-English language skills to obtain preoperative informed consent from LEP patients.

"Additionally, the method of language assistance the surgeon selected in the hypothetical scenarios presented was noted. The options included using one's own non-English language skills, calling a professional interpreter, using an ad-hoc (untrained) staff member, and, in situations where the patient was accompanied by family, the patient's adult family member or child." (Patel, 2016, p. 515)

How? #ⓘ

A 32-item anonymous online (REDCap) self-report survey covering demographics, training, non-English language skills and their clinical use, and hypothetical LEP consent scenarios; surgeons self-rated fluency on an adapted Interagency Language Roundtable (ILR) scale, with "good," "fair," or "poor" classified as nonfluent.

"Participants were considered fluent if they self rated as 'excellent' or 'very good' and nonfluent if they self rated as 'good', 'fair,' or 'poor'." (Patel, 2016, p. 515)

Who? #ⓘ

The subset of surveyed surgeons who reported using their non-English language skills with patients (n = 48), drawn from 158 respondents (33.1% response rate, 158 of 477) across all surgical specialties at a single academic medical center in Boston, MA.

"Table 2 looks specifically at the experience of surgeons who use their non-English language skills with patients (n = 48)." (Patel, 2016, p. 517)

Other Notes #

The paper distinguishes fluency by self-rated ILR level: fluent speakers reported using their skills more often than nonfluent speakers (P < 0.001), but nonfluent speakers also reported using their skills at least some of the time (Patel, 2016, p. 517).

Caveats #

• Findings are surgeon self-report subject to selection bias so reported behavior may diverge from actual consent practice All findings are surgeons' self-reported behaviors, not directly observed consent encounters, and the sample may have been subject to selection bias favoring respondents with a particular interest in the LEP patient population. Such respondents may have given answers aligned with institutional policy rather than their actual practice, so the reported use of ad-hoc interpreters and own-language skills could differ from real behavior. The authors argue this bias likely leads to an understatement of the problem, since respondents still reported policy-contravening behavior (e.g., using minors).
• A single urban academic center with a low response rate and small sample limits representativeness of the surgeon survey The survey had a low response rate (33.1%) and small sample, limiting the strength of associations, and was conducted at a single urban academic medical center, which may not represent national physician and patient demographics or the behavior of surgeons at institutions with different interpreter resources and policies. Prevalence estimates and fluency subgroup comparisons should therefore be generalized cautiously.