An observed higher elective-arthroplasty rate among LEP patients does not by itself establish a language-driven treatment effect given selection bias and disease-severity confounding

An observed higher elective total joint arthroplasty rate among patients with limited English proficiency does not by itself establish a language-driven treatment effect: the association is vulnerable to selection and referral bias (single-center capture can understate the comparison group's surgery rate) and to confounding by greater baseline disease severity (a higher proportion of LEP patients presenting with end-stage arthritis), and it cannot be causally interpreted without functional/patient-reported outcome measures and data on interpreter type and quality.

Provenance (reviewer flag). This claim is the central methodological argument of a Letter to the Editor by Zirvecan Güneş, MD (J Arthroplasty 41 (2026) e20–e21) commenting on the Angeles et al. study; it is a commentary with no primary data of its own. It bears on (and cautions the causal reading of) the descriptive claim C-0115ClaimC-0115Initial AI draftPatients with limited English proficiency are more likely to receive elective total joint arthroplasty than English-proficient patientsPatients who have limited English proficiency (LEP) are more likely to receive elective total joint arthroplasty than fully English-proficient (FEP) patients — an unadjusted 46% versus 38% surgery rate and roughly 20% hi….

Grounding quotes (from the Güneş commentary letter) #

"The single-center design captures only surgeries performed within the authors' institution. The actual total joint arthroplasty rate for FEP patients could be higher than reported because they might choose to get second opinions or undergo surgery at different facilities, which would make the observed difference between LEP and FEP cohorts appear larger than it actually is." (Güneş, 2026, p. e20)

"That the vast majority of LEP patients had Kellgren—Lawrence grade 4 arthritis compared to FEP patients (71 versus 55%) probably also contributed to this surgery rate. However, in the absence of patient-reported outcome measures (PROMs), such as pain or function scores, it is impossible to discern whether the remaining 20% adjusted odds reflect a lower symptom burden of disease or preference or different communication dynamics." (Güneş, 2026, p. e20)

"It appears that no information on the type of interpreter is provided, and nothing is said about the quality of the interpretation received. This makes it quite impossible for readers to know whether surgical preference affected patient decision or merely restricted access to nonoperative options due to lack of communicative ability." (Güneş, 2026, p. e20)