The objective of the proposed research is to enable real-time monitoring of respiratory health in clinical settings via wearable-based multimodal physiological signals. This requires two main components: robust instrumentation and bio-signal processing. Specifically, respiratory status will be inferred from respiratory sounds, transthoracic bioimpedance, and chest kinematics, with an emphasis on Lung Sounds (LS)  and Impedance Pneumography (IP).  In our first completed contribution, we adapted for respiratory monitoring a wearable multimodal sensing suite previously developed by our group for knee health assessment and validated its feasibility for continuous respiratory monitoring in clinical settings. The system demonstrated its capability for tacking changes in pulmonary fluid surrogates, LS signatures, and respiratory markers in patients hospitalized with Heart Failure (HF), wherein respiratory distress is the hallmark. Our proposed next step is to elucidate novel physiological markers that leverage the cross-integration of these wearable-based signals to predict clinical status in pediatric patients hospitalized with an acute asthma attack. Although markers derived from these signals separately have been widely used, their cross-integration remained to be investigated. Thus, our second completed contribution was to develop bio-signal processing techniques to extract clinically-relevant physiological markers that fuse these sensing modalities. We demonstrated that using IP to provide breathing-phase context enhanced the predictive power of LS. We propose furthering these efforts by modeling clinical status using interpretable machine learning models trained on a holistic and robust set of physiology-driven respiratory markers, as well as clinical data from the electronic medical records. This will result in a novel and comprehensive evaluation of the usability of these wearable-based signals for continuous respiratory status assessment, thereby laying the foundations for a real-time monitoring system.