DefinePK

Abstract

This paper aims to develop a concept of utilizing an AI-based predictive model to address why early signs of patient condition worsening go unnoticed in nursing care. It was also found that most of the monitoring activities employed in the hospital context do not provide an early warning sign for a poor outcome for the patient and consequent increased workload of the nurse. In the conduct of the study, the authors used both quantitative and qualitative techniques, and a new RNN model was developed and implemented using electronic health records for developing deterioration risk assessments. Causal analysis supported the probability of the model to predict outcomes in other patients while descriptive analysis described the perceived easy use of the model as adopted by the nursing staff and the specific ways and times when it would be used in their working practice. The progressed model yielded an average ROC-AUC of 0.89 for the detection of clinical cases with a parallel decrease of response time for clinical alert by 15 % enabling the nurses to focus more on high-risk patients. The subjective beneficial results obtained from the qualitative feedback in the descriptive part include the ability of the model to enhance objectives, situational awareness, and decision making though this was attributed to some occurrences that were found to give out false positives and therefore required some enhancements. From the above discovery, the argument may be made that AI improved patient care outcomes due to the capability of the system to deliver evidence-based proactive nursing interventions and reduce instances of preventable adverse outcomes. Thus, it is proposed that the model be adopted in several clinical projects to study the chronic impact on the final patient outcome. Therefore, it can be argued this study contributes to the literature on knowledge enhancement on the use of AI tools in enhancing nursing practice to enhance quality and safety within acute healthcare facilities.