Artificial Intelligence Inpatient Deterioration Forecasting Under Variable Nurse Charting Interval Irregularity: An Irregular-Time Hawkes Process Evaluation of Healthcare Rapid-Response Activation

Nurse charting intervals in inpatient wards are irregular, and the irregularity is informative: deteriorating patients are charted more often, so the observation times of a deterioration-forecasting model depend on the outcome the model is meant to predict. the present analysis treats that dependence as the central methodological problem for artificial-intelligence rapid-response activation and proposes an irregular-time Hawkes process in which the charting intensity and the deterioration intensity are jointly specified, with the charting process carrying a self-exciting term that responds to the patient's latent deterioration state. The proposal is evaluated as a design study rather than an empirical trial: no observational data are used or claimed, and the operating characteristics of the before-after-control-impact estimator that would be used to evaluate a deployed system are characterised under a declared variance structure with between-cluster SD = 6.0, residual SD = 4.5 and intraclass correlation = 0.64, across 2000 replications per design cell for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points. The design analysis indicates that 54 clusters are required to detect a 5-percentage-point effect at 80% power, a requirement that exceeds every cluster count in the evaluated grid. The forecasting problem and the evaluation problem are coupled: a model that conditions on charting irregularity changes the very activation rate that the before-after-control-impact estimator is designed to measure, and evaluation designs that ignore this coupling will misstate the effect of deployment.