A shift of perspective on system safety has been made from failure of elements or errors to performance variability to understand events occurring in modern complex socio-technical systems. While the Functional Resonance Analysis Method (FRAM) is a well-established methodology for analyzing system behaviors in terms of variability, assessing relative risk levels requires more specified representation and handling of the quantitative aspects of variabilities. This dissertation proposed and examined a quantitative scheme to use the FRAM by defining rules for variability propagation and aggregation in consideration of interactions and adaptive operation of functions. Quantitative value calibration for the variability is proposed to consider the structure of FRAM model using quantitative variability propagation. Potential variability of each function can be characterized based on the calibration result with expert evaluation of the performance margin of each function. Then, the quantitative FRAM analysis is suggested for comparative analysis of potential risks and system changes. Quantitative variability simulation is also proposed by defining system reactions to events in terms of variability which allows to examine organizational factors of the system. Application to an emergency response system for infectious disease in relation to managing crisis response for 2019 coronavirus disease (COVID-19) demonstrates the usefulness of the proposed method for assessing the relative importance of potential risks and critical conditions. Although the example was taken in the infectious disease containment case, the proposed method can generally support strategic decision-making from a safety perspective in other domains as well.