Aims/hypothesis <p>The study aimed to model the progression of diabetic kidney disease (DKD) in a multi-ethnic Asian population and evaluate the economic impact of a risk-stratified intervention for DKD management.</p> Methods <p>Using the Singapore Diabetic Cohort (<i>N</i>=9660), multi-state Markov models were constructed to derive chronic kidney disease (CKD) stage-specific transition probabilities for disease progression based on Kidney Disease: Improving Global Outcomes (KDIGO) guidelines. These were internally validated in 50% training and 50% testing datasets and further validated in national administrative datasets (<i>N</i>=73,256). A cost-effectiveness and budget impact analysis for a hypothetical intervention of sodium–glucose cotransporter 2 inhibitors (SGLT2is) was conducted from a healthcare system perspective, across different time horizons, in low-risk (G0–G2) and high-risk individuals (G3–G4) based on eGFR-defined DKD risk stages.</p> Results <p>Annually, more than half of the participants (53.0–88.8%) stayed in the same DKD stage, with the probabilities of progressing or regressing to the next DKD stage being highest between stages G0 and G1. The SGLT2i intervention was cost-effective when given in low-risk, but not in high-risk, individuals with type 2 diabetes in both the short and the long term. Additionally, implementing the SGLT2i intervention would incur a 5 year cumulative cost increase of 41.5%, assuming an annual 5% increase in market uptake.</p> Conclusions/interpretation <p>We developed a robust model of DKD progression based on real-world data and demonstrated its value through economic modelling of risk-stratified interventions with known effectiveness. This model can aid policymakers and health professionals in evaluating population-level interventions for evidence-based policy decisions.</p> Graphical Abstract <p></p>