Translated consent documents (CDs) and use in non-industry sponsored studies (NISS).

Abstract

6553 Background: Patients (pts) from historically underrepresented groups often have limited English proficiency (LEP). One challenge specific to obtaining consent from these pts to participate in clinical trials is the need for translated CDs. CD translation leads to enrollment delays and increases costs to study sponsors. We hypothesized that NISS, for which the principal investigator pays translation costs, would consent proportionately fewer pts with a primary language other than English and pts with LEP compared to industry sponsored studies (ISS). Methods: Pts’ primary language and English proficiency were assessed for all 12,082 consent events with appropriate available data at the Jonsson Comprehensive Cancer Center from 2013 to 2018. Pts with LEP were pts (or parent/guardian for pediatric pts) who had a primary language other than English, and the electronic health record indicated that the pt required a translator within 6 months of the consent date. CD language was documented per chart review when available, and when not, languages of available IRB-approved CDs were evaluated. The proportion of consent events for pts with a primary language other than English and with LEP were compared between NISS and ISS using generalized estimating equations, and the odds of signing CDs in the pts primary language were evaluated by multivariable analyses. Discrepant CD language data was evaluated by McNemar’s test. All analyses were performed using SAS 9 (SAS Institute, Cary, NC, USA). Results: Pts with a primary language other than English represented 8.1% of consent events in ISS vs 4.4% in NISS (p<0.001) and 5.5% vs 2.8% (p<0.001) for pts with LEP. Pts with a primary language other than English represented 4.5% vs 1.2% of consent events utilizing CDs in the pts’ primary language in ISS vs NISS (p<0.001) and 3.7% vs 0.9% (p<0.001) for pts with LEP. On multivariable analyses, the odds for pts with a primary language other than English of signing CDs to a NISS was 0.74 (CI, 0.63-0.94, p=0.005), and of signing CDs in their primary language was 0.38 (CI, 0.27-0.52, p<0.001) compared to pts whose primary language was English. The odds for pts with LEP of signing CD to a NISS was 0.74 (CI, 0.58-0.95, p=0.02), and signing CDs in their primary language was 0.35 (CI, 0.25-0.50, p<0.001) compared to pts whose primary language was English. Of 52 pts who signed CDs for both ≥1 NISS and ≥1 ISS, 18 signed in discrepant languages, with 16 of those 18 signing translated CDs for the ISS but not the NISS (p=0.002). Conclusions: Although the role of cost as an effect driver cannot be proven in a retrospective analysis, consistent with our hypothesis, the proportion of consent events utilizing appropriately translated CDs for pts with a primary language other than English and pts LEP was lower in NISS vs ISS, driven by lower odds of signing CDs and lower frequency of signing translated CDs. Approaches that reduce translation costs for investigators should be investigated.