Early detection through proper screening and identification of high-risk individuals is essential to help decrease the morbidity and mortality in Diabetes Mellitus. Hence, this analytical cross-sectional study aimed to determine the performance of the Recalibrated FINDRISC Diabetes Mellitus risk screening tool as applied in the study setting of selected Municipalities of Zamboanga del Sur. It specifically aimed: first, to evaluate the risk screening tool’s (1) sensitivity, (2) specificity, (3) predictive values, (4) likelihood ratios and (5) overall performance through the ROC, AUC and Q-value; and second, to compare the performance of the three screening tools, namely: the (1) Recalibrated FINDRISC, (2) FINDRISC, and the (3) Modified FINDRISC. A total of 1169 respondents (162 previously known diabetic, and 1007 undiagnosed individuals) were chosen through convenient sampling. The Diabetes Mellitus Risk Score (either positive or negative) by each tool is confirmed by a fasting blood sugar test (reference test). Two-by-two tables were then generated to calculate sensitivity, specificity, predictive values, and likelihood ratios and the corresponding ROC curves, AUC, and Q-values. Results showed that the FINDRISC has the highest sensitivity (96%) and the Recalibrated FINDRISC has the highest specificity (84%). Looking through each tool’s overall performance, the Recalibrated FINDRISC’s results showed that it has unsatisfactory performance both in the AUC = 0.61 (95%CI: 0.57-0.64) which is between the ‘sufficient’(0.6-0.7) and ‘bad’ (0.5-0.6) performance level in terms of its diagnostic accuracy; as well as in its Q-value of 0.55, which is interpreted as having ‘bad’ performance in terms of its discriminatory ability as a risk screening tool; thus it is not recommended for use in populations having similar characteristics to the study. On the other hand, both the FINDRISC (has the best Q-value = 0.81) and Modified FINDRISC (has the best AUC value = 0.83 [95%CI: 0.81-0.86]) have satisfactory performance hence it could be utilized and maybe usefully applied to identify high-risk individuals in populations with characteristics similar to the study.

Keyword: Recalibrated finnish diabetes risk score Type 2 diabetes mellitus