Processing

• Filtering, resampling and handling irregular and missing observations
• Feature extraction, including heart-rate variability and waveform morphology
• Windowing and label alignment for event prediction

Prediction

• Early warning scores as the baseline any model must beat
• Sepsis and deterioration prediction, and the published failures worth studying
• Lead time versus precision, and the ward capacity to respond

Alarms

• Alarm fatigue as the dominant clinical failure mode of monitoring systems
• Alarm suppression, escalation logic and the safety case for silencing anything
• Human factors in the design of a bedside or remote monitoring display

Remote Care

• Remote monitoring and hospital-at-home programme design
• Connectivity, device management and data governance outside the hospital
• Evaluating whether monitoring changed outcomes rather than just generating data