Evaluating the performance of TAG-IT for prediabetes detection in Indonesian population

Abstract

Early detection of prediabetes plays a critical role in preventing type 2 diabetes mellitus (T2DM), especially within primary care, where access to laboratory testing may be constrained. Non-laboratory-based risk assessment instruments, including the tool to assess the likelihood of fasting glucose impairment (TAG-IT), can facilitate preliminary risk screening. This study sought to determine the diagnostic accuracy of the TAG-IT questionnaire in detecting prediabetes, using the oral glucose tolerance test (OGTT) as the reference standard. A cross-sectional design was implemented across three community health centers in the Special Region of Yogyakarta, Indonesia. Although 308 individuals were initially enrolled, only 93 participants with complete datasets were eligible for final analysis. The discriminative capacity of TAG-IT was evaluated through receiver operating characteristic (ROC) curve analysis along with a contingency table. Among the participants analyzed, 24.7% (23/93) were classified as having prediabetes. The TAG-IT tool exhibited modest discriminatory performance, yielding an AUC of 0.656 (95% CI 0.525-0.786; p = 0.026). Using the identified optimal threshold, sensitivity reached 52.2% while specificity was 67.1%. The negative predictive value was 81.0%, indicating better performance in excluding low-risk individuals. Overall, TAG-IT demonstrated moderate utility as a preliminary screening instrument in primary healthcare, particularly for identifying individuals unlikely to have prediabetes.