Seasonal Autoregressive Moving Averages and Crowd Wisdom Models in Medical Forecasting in China : Evidence from Affiliated University Hospitals

Abstract

The insight that crowd responses to an estimation task can be modeled as a sample from a probability distribution has invited comparisons with individual cognition and conventional decision making tools. However, within the past decade, conflicting evidence of the generalisability of crowd wisdom techniques in forecasting has reignited the debate on their robusticity in healthcare decision making. Patient arrivals at otorhinolaryngology departments of selected hospitals in China were modeled and predicted using a prediction market crowd wisdom technique. The mean average percentage error was compared with similar predictions using the seasonal autoregressive integrated moving averages. The prediction market quarterly patient flow prediction defeated the seasonal autoregressive moving averages model by 0.06% in the affiliated hospital of the Guilin medical university and by 0.56% in the affiliated Jiangsu university hospital but were weaker in terms of daily and monthly predictions. Moreover, we observed a striking seasonal variation in otorhinolaryngology department attendance; an aberration from past knowledge that inspires curiosity for further research