44 - Audit of Translated Consent Documents Validates Current Translation Requirements

Background: : Our IRB evaluates translation methods and translator qualifications when reviewing translated consent forms. Researchers submit a translation attestation but we do not have a mechanism to assess translation accuracy. To gauge the efficacy of our approach, we conducted an audit of translated consent forms.

Program

Description: : In order to assess overall translation accuracy and identify whether there are differences in accuracy between professional translation services and non-professional translation, our Post-Approval Monitors (PAMs) conducted an audit of IRB approved translated consent materials. They identified studies with IRB approved translated consent forms and selected 41 studies from 11 departments with 19 languages represented. This generated 48 unique consent documents for audit. A professional translation service provided back-translations of the selected consent forms for comparative analysis against the IRB-approved English versions.

Program

Assessment: : The PAMs compared the back-translated consent forms to the IRB approved English versions to identify language discrepancies. Discrepancies were analyzed and categorized as: cultural differences, translator errors, version errors, and intentional changes. PAMs utilized Google Translate as a secondary tool to assess discrepancies between originals and back-translations and conducted a root-cause analysis for any discrepancies. When they were not able to verify a discrepancy using Google Translate, PAMs contacted investigators for clarification. Discrepancies were assessed by PAMs for possible noncompliance.

Of the 48 consent forms reviewed, 30 consents had no issues with the translations (62.5%), 18 consents had observations that were determined to be minor noncompliance (37.5%). There were no consents with observations considered serious noncompliance. Nine instances of minor noncompliance were due to translator error. Other instances included translating the wrong document and failing to inform our IRB after modifying with the local review board.

Most discrepancies did not change the meaning of the IRB approved English equivalents. These types of discrepancies were not determined to be noncompliance.

Limitations:: There is no way to search our database for studies using translations; thus, our sample does not encompass all translated materials used. Some languages had no direct equivalents for certain words. An English consent form might describe blood collection of 6 tablespoons, which is an imperial system for measurement, but in some cultures/languages, a formal measurement system is not used. In some cases, blood collection measures were described in generic terms (e.g., 6 small spoons).

Discussion: : Most consent forms had no issues or minor issues. No serious noncompliance was identified. There were no appreciable differences in the error rate between consent forms that were translated by a professional vs. non-professional translator where the error was attributed to the translator. This suggests that, generally, our current policy requirements regarding translated consent forms result in accurate translations.

Our recommendation is for IRBs to allow flexibility in translation policies for translations from qualified non-professional, as this reduces burden on both the researcher and the IRB.