Preferred language was self-reported by the patient at 68 percent of BMT centers

Description #

Across all geographic regions, respondents indicated that a patient's preferred language was most commonly identified through patient self-report (68%). This describes how transplant centers ascertain language preference — the front-end step in providing language-concordant services.

"Across all geographic regions, respondents indicated that preferred language is self-reported by the patient (68%)." (Idossa, 2019, p. 1033)

Methods Context #

What? #ⓘ

The observable: how the transplant center identifies a patient's preferred/needed language (e.g., patient self-report vs. staff documentation).

"How TCs identify language needs" (Idossa, 2019, p. 1032, Fig. 1)

How? #ⓘ

A national cross-sectional, web-based survey of transplant-center staff, with the survey's second domain covering how centers identify language needs.

"This cross-sectional, web-based survey was comprised of 30 items within five domains (Fig. 1): ... (2) how TCs identify language needs." (Idossa, 2019, p. 1032)

Who? #ⓘ

The 82 transplant-center-staff respondents (59% response rate) across all U.S. census regions.

"The survey yielded a 59% response rate (N = 82)." (Idossa, 2019, p. 1033)

Other Notes #

Self-report was the dominant identification method even as system-level gaps in tracking and documentation persisted (see the system-level barrier EVD).

Caveats #

• Cross-sectional TC-staff survey (59 percent response) underrepresents small-volume centers and non-social-worker roles with high item missingness The estimates come from a cross-sectional web-based survey with a 59% response rate in which not every transplant center provided input; patient educators/advocates and transplant coordinators were poorly represented (respondents were overwhelmingly social workers), and small-volume centers (fewer than 100 BMTs/year) were underrepresented relative to medium and large centers. In addition, many items carried a relatively high percentage of missing responses (e.g., 16.5–22.8% missing on Table 3 items), which the authors attribute to respondents not knowing or the item not applying. Together these features limit the representativeness and precision of the reported frequencies of needs and barriers.