Dear Editor

I read with interest the study by Almonte-Becerril and colleagues, which estimated the diagnostic accuracy of the FINDRISC (Finnish Diabetes Risk Score) questionnaire compared with glycated hemoglobin (HbA1c) in 148 adults living in Ixtepec and Huehuetla, Puebla [1]. This study represents a valuable contribution to the scientific community, as it documents, for the first time, the performance of this questionnaire administered to an Indigenous Mexican population. The study was reported in accordance with STARD [2], with assessor blinding and linguistic mediation through verbal explanations provided by bilingual personnel.

However, I would like to point out that the limitations section does not address the quality of the reference standard. In this study, it is important to distinguish between the reference condition used and a confirmed clinical diagnosis. Any participant with a single HbA1c determination equal to or greater than 6.5%, obtained with a point-of-care device (Eclipse A1c), was classified as "compatible with T2DM," without any claim that a definitive clinical diagnosis had been established. This observation is not intended to question the authors' interpretation of their results, but rather to point out a methodological risk: the American Diabetes Association Standards of Care in effect for 2025 [3]—the reference point used by the authors for the HbA1c cutoff value—clearly state that HbA1c testing for diagnostic purposes must be performed exclusively in laboratories certified by the National Glycohemoglobin Standardization Program (NGSP), and that, in the absence of unequivocal hyperglycemia, the diagnosis requires a second abnormal result on the same test, measured either simultaneously or at a different time point, or an abnormal result on a different test (fasting plasma glucose) [3]. In addition, the NGSP itself explicitly notes that point-of-care devices are not recommended for diagnosis [4]. Although the authors reported daily calibration of the Eclipse A1c device, they did not report its diagnostic traceability or confirmation of its results. For this reason, the reference variable used as "compatible with T2DM" may have incorporated some degree of misclassification. Such misclassification of the reference standard could bias the study’s estimates of sensitivity and specificity.

Notwithstanding the above, this does not invalidate the important contribution the study makes regarding the performance of the FINDRISC questionnaire in communities with limited access to confirmatory testing. It emerges as a useful initial screening tool, with moderate discrimination, low specificity, and a need for subsequent confirmatory testing. Future diagnostic accuracy studies should implement and describe a predefined confirmatory procedure to help reduce potential misclassification.