Healthcare facilities globally face challenges in managing patient inflow and maximizing the utilization of resources. Incorrect patient flow predictions have the potential to result in overcapacity, waiting lists, staff burnout, and ineffective care delivery. The paper undertakes a systematic review and critical analysis of advanced predictive models utilized in patient flow prediction in clinics and hospitals. Through a critical examination of various modeling techniques, including time-series forecasting, queuing theory, discrete-event simulation, and machine learning algorithms, this paper identifies their strengths, limitations, and areas for further improvement. The study brings to the fore the possibility of using artificial intelligence (AI), through deep learning and ensemble modeling, to increase forecasting accuracy. Empirical evidence from hospital pilot study reports, case reports, and peer-reviewed journals suggests that while traditional models offer initial perspectives, AI-driven models offer greater flexibility to adapt to system uncertainty and real-time information. This paper suggests implementing dynamic predictive tools within hospital management systems to support strategic planning and operational planning, especially in emergency departments (EDs), intensive care units (ICUs), and outpatient clinics.
Oware et al. (Wed,) studied this question.