Regime-based model of sustainability of a regional anti-drug system.

<jats:p>The assessment of sustainability of regional anti-drug systems is traditionally based on monitoring clinical and resource indicators, which limits the timely detection of latent regime shifts and loss of returnability. Objective: to develop and empirically validate a computable regime-based model of sustainability of a regional anti-drug system based on the reconstruction of latent adaptability. Materials and methods. Data from the medical subsystem of the Kabardino-Balkarian Republic for 2014–2024 were analyzed. Min–max normalization was applied to diagnostic indicators. An integrated medical index (M_d*) was constructed. Latent adaptability (At) was modeled using a recurrent equation with an inertia parameter (λ). Calibration of λ was performed using rollingorigin validation with minimization of the mean absolute error (MAE). The boundaries of the admissible regime domain (Ω_stable) were determined through retrospective analysis and a quantile-based criterion. Results. A decline in latent adaptability At from 0.9439 to 0.6863 (–27.3%) during 2014–2024 was identified, despite substantial variability in the observed medical index M_d*. The optimal inertia parameter λ was 0.05, indicating high institutional inertia of the system. Scenario modeling for 2025–2029 demonstrated that inertial and compensatory strategies do not restore the system’s membership in the Ω_stable domain, whereas the adaptive scenario ensures a sustained reversal of the At trajectory. Conclusion. A computable regime-based sustainability model is proposed, enabling the early diagnosis of loss of returnability prior to clinical crisis manifestation and supporting the formation of managerial early-warning triggers. Integration of At into the AURA-ML digital analytical framework expands the strategic management toolkit for regional anti-drug systems.</jats:p>