Professionally translated bilingual consent forms improve understanding of treatment intent in simulated systemic anti-cancer therapy (SACT) consultations, whereas machine translation introduced more critical errors, UK researchers have reported.

Limited English-language proficiency can be a significant barrier to equitable cancer care, particularly in discussions about SACT, where understanding treatment intent and risks is essential. However, patient information and informed consent documentation are often available only in English.

A randomised UK study, published in the journal Supportive Care in Cancer, therefore assessed whether different translation approaches could improve understanding of SACT information among adults with limited English language proficiency.

The study recruited 123 healthy volunteers from Bengali- and Sylheti-speaking communities in London who could read written Bengali and self-identified as preferring medical information in Bengali or Sylheti rather than English.

Participants received either a professionally translated Bengali myeloma information booklet or a Google Translate-generated version.

SACT consent consultations

One to two weeks later, participants underwent a simulated SACT consent consultation with a haematologist and an interpreter and were then further randomised to receive either a standard English-only consent form or a professionally translated bilingual English–Bengali version.

The primary endpoint was accurate understanding of treatment intent, specifically, recognising that myeloma treatment was not curative but aimed at prolonging survival and improving quality of life.

Among 121 participants included in the booklet analysis, 15.7% reached the primary endpoint and correctly understood SACT treatment intent after reading the translated information, with no significant difference between professional and machine-translated versions.

However, an independent translation assessment showed poorer quality in the machine-translated booklet. Using the Chartered Institute of Linguists’ criteria, the machine translation scored 51/100 compared with 73/100 for the professional translation.

There were 11 critical meaning-changing errors and 19 additional accuracy issues in the machine-translated version, compared with one critical error and four accuracy issues in the professional translation.

Bilingual forms and understanding treatment intent

A total of 91 participants completed the simulated SACT consent consultation component. Correct understanding of treatment intent was achieved by 60% of participants who received the bilingual consent form, compared with 34.8% of those who received an English-only form.

Randomisation to the bilingual form significantly increased the likelihood of correctly understanding treatment intent. Participants who received bilingual forms also reported greater confidence in understanding the consultation and in answering treatment-related questions from family members.

Bilingual consent forms may improve comprehension by allowing patients to read translated information directly and by supporting interpreters in using consistent terminology. However, the approach was not universally effective, with 40% of participants in the bilingual group still failing to correctly identify treatment intent.

The study used only healthy volunteers rather than patients with cancer, which limits the findings, as did having simulated rather than real-world consultations and immediate post-intervention assessment rather than longer-term recall. The exploratory study was also not formally powered to assess definitive efficacy outcomes, the authors noted.

Despite these constraints, they concluded that bilingual interlinear consent forms for SACT could offer a simple, cost-effective way to improve communication and health equity for patients with limited English proficiency.

However, as most participants still failed to understand crucial treatment information regardless of translation type, the authors also called for alternative communication approaches, including audio and video resources, and cautioned against unsupervised machine translation without human review.

Reference
Hibbs SP et al. Translation approaches to support systemic anti-cancer therapy consent for individuals with limited English proficiency. Support Care Cancer 2026;34:317.