Linguistic validation of Functional Assessment of Cancer Therapy–Cognitive Function (FACT-Cog): methodological concerns

Abstract

Dear Editors: We read the article written by Joly et al. [1] with great interest. The authors conducted a validation study on the French-translated Functional Assessment of Cancer Therapy–Cognitive Function (FACT-Cog) version 3 [2]. It was concluded that the translated instrument is reliable for the self-reporting of cognitive abilities in the French population of cancer patients. The study was well conducted with a number of strengths worth emphasizing. First, the methodology incorporated both qualitative and quantitative approaches to ensure accurate transfer of meanings from the English to French translation. Refinements were then made to the French version after the pretesting phase to improve comprehensibility of the items that were originally difficult to understand. Moreover, this study involved a heterogeneous sample of patients with different cancer conditions, cancer treatment statuses, and ages to ensure generalizability of the results. Reliability of the French version of FACT-Cog was also assessed through Cronbach’s alpha and item–total correlation. Currently, there is an evident lack of published data on self-reporting psychometric assessments of cognition within the cancer population [3]. To effectively evaluate cognitive changes in multiethnic cancer patients, we have also highlighted the necessity to adopt a culturally relevant approach [4]. Hence, rigorous linguistic validation and cultural adaptation conducted by a bilingual or bicultural research team is recommended to attain both conceptual and linguistic equivalence of a translated neuropsychological tool [5, 6]. Although Joly et al. had undergone a vigorous translation process to ensure content and semantic equivalence of the translated tool, it is imperative to conduct the comparability and equivalence between the English and French versions of the FACT-Cog, as there is no assurance that the translated tool is equivalent to the original English version unless the comparability in the psychometric properties of the two tools is being performed. One study had shown that the original English EORTC QLQ-C30 and its nine European translations found indications of nonequivalence in seven of nine translations and at least two of the versions presented with linguistic problems in the translation [7]. Measurement equivalence is necessary to facilitate the pooling of results from multinational clinical trials and identify problematic items that differ in semantic structures from the original language version [6]. In addition, as the authors have graciously pointed out, the cross-sectional design of their study did not allow the evaluation of procedures including test–retest reliability and responsiveness to cognitive changes within their patients. Other important components of the validation should involve the evaluation of convergent, discriminant, and known-group validities of the translated FACT-Cog, as one may question whether the tool is accurately evaluating the desired constructs of cognitive functioning. We have recently conducted a validation study of the simplified Chinese version of FACT-Cog within an Asian population of cancer patients [8]. During our cognitive debrief phase with 30 Chinese-speaking patients, our Y. T. Cheung :A. Chan (*) Department of Pharmacy, Faculty of Science, National University of Singapore, Block S4, 18 Science Drive 4, Singapore 117543, Singapore e-mail: phaac@nus.edu.sg