In the current work, we present our position towards the modeling of a decision support system for health assessment and prevention of risky situations in the life of seniors. A two-layer architecture based on the Fuzzy Cognitive Maps (FCMs) methodology is considered as the main tool to encode medical knowledge and reason about the health status of seniors. This is the initial step, since outputs of several discrete FCM models are then combined to form the input concepts for an upper level schema (the health status of the senior) of upper-level concepts and relationships, so as to infer about recommended individualized therapies and interventions that would help therapists to prevent risky situations and enhance the independence and quality of life of the seniors. A modeling sample of the depression assessment is presented at the first level of the proposed architecture, while am integration schema of the proposed approach provides a view of our future extended model.