The prevalence of type 2 diabetes mellitus is rising, especially in low- and middle-income countries, where the situation is worsened because around half of cases are unaware of the disease. Universal screening utilizing blood markers can be challenging in resource-constrained settings. The identification of these individuals can be potentially addressed using risk scores and neuropathy assessment tools. This study aimed to assess the diagnostic accuracy of the FINDRISC, a blood-free risk score, three neuropathy assessment tools (EZSCAN, pupillometer, and biothesiometer), alone and in combination. A population-based study was conducted enrolling a sex-stratified random sample of participants from Tumbes, a semiurban area in the north of Peru. Undiagnosed T2DM was the outcome, defined using WHO OGTT thresholds. Diagnostic accuracy of the FINDRISC and neuropathy tools was evaluated using the area under the ROC curve (aROC) and respective 95% confidence intervals (95%CI). Data from 1609 participants were analysed, mean age 48.2 (SD: 10.6) years, 810 (50.3%) females. A total of 176 (10.9%) individuals had T2DM, and only 71 (4.7%) had undiagnosed T2DM. The diagnostic accuracy of the FINDRISC was aROC = 0.69 (95% CI: 0.64–0.74), with a sensitivity of 69% and specificity of 67%. Among devices, the EZSCAN (aROC = 0.59; 95%CI: 0.53–0.66; sensitivity of 59% and specificity of 54%) and biothesiometer in the third metatarsal head (aROC = 0.60; 95%CI: 0.53–0.67; sensitivity of 31% and specificity of 85%) performed best. A combination of the FINDRISC and the biothesiometer had the best diagnostic accuracy, with a similar aROC of FINDRISC alone (AROC = 0.69; 95%CI: 0.68–0.78), with a sensitivity of 79% and a specificity of 59%. Our results confirm that combination of the FINDRISC and biothesiometer can improve diagnostic accuracy of the FINDRISC and biothesiometer alone, increasing sensitivity without affecting specificity or the area under the ROC curve.