Stakeholder perspectives on implementing ED time series forecasting: a qualitative interview study.

backgroundEmergency departments (EDs) around the globe are experiencing increasing patient attendance and staffing challenges, making it difficult for management to balance supply and demand. Time series forecasting (TSF) is a statistical technique that uses historical data and external variables to predict phenomena over time. Given the growing interest in TSF applications in emergency care, the aim of this study was to explore how these technologies could be aligned with everyday clinical demands to close the translational science gap.

methodsThis was a prospective qualitative interview study. Through interviews with European emergency department directors, we collected and analysed perspectives on the practical applicability of time series forecasting in everyday emergency care. The study is reported in accordance with the Consolidated Criteria for Reporting Qualitative Research (COREQ).

resultsWe interviewed 26 ED directors in 19 countries. Over half had prior knowledge of time series forecasting technology. Only three countries had TSF implemented in daily practice. Most ED directors reported that their management decisions are informed by data and indicated openness to the adoption of additional decision-support tools, such as TSF. Novel, previously undescribed external variables were identified, including local primary care availability, fruit harvesting, and neighbouring hospital conditions (e.g. ED closures).

conclusionsWe recommend that future TSF modelling studies in ED implement, at a minimum, both short- and long-term forecast horizons; conduct in-depth interviews with local senior staff to align models with operational needs; and validate models using real-world data from deployment sites. Through semi-structured interviews, we identified key directions for further TSF model development to enhance usability and support translation from academia into clinical practice. STUDY REGISTRATION: This study was preregistered via Open Science Framework site ( https://doi.org/10.17605/OSF.IO/SGZH7 ).