Accurate battery state estimation is critical for enabling intelligent operation and predictive maintenance. However, many existing approaches rely solely on electrochemical models without explicitly incorporating degradation or lifetime models. Joint- and dual-estimator architectures have demonstrated the ability to track slowly evolving health states, such as capacity, alongside faster dynamic states. While useful and effective, these capacity estimates can be noisy and are not directly suited for prognostic analysis as they do not provide a framework for predicting future degradation or remaining useful life. As an alternative, degradation assessment is also commonly performed using diagnostic protocols that enable targeted measurements, though these protocols require time and interrupt normal operation, limiting their frequency.