The effect of visibility on forecast and inventory management performance during the COVID-19 pandemic

Abstract

During the COVID-19 pandemic, healthcare organizations suffered a shortage of essential medical supplies, such as personal protective equipment, which resulted in severe consequences. This study aims to assess the impact of one potential factor for this shortage, i.e., the lack of visibility over the consumption of personal protective equipment. To do so, different forecasting methods combined with a periodic review inventory system are tested on semi-simulated data that include various visibility issues. The forecasting methods are categorized based on the data used. The Holt and naïve methods are selected as demand-based forecasting methods, and a modified compartmental epidemiological model is explored for its use of pandemic data to forecast demand. This paper studies three of the most common data visibility problems. Specific scenarios have been developed to analyze the impact of (1) delayed data, (2) temporally aggregated data, and (3) erroneous data on the performance of the system. Our findings indicate that, in most cases, data visibility issues directly influence the healthcare supply chain and diminish the performance of the system. However, when these visibility issues result in exponentially large over-forecasts, we observe a performance improvement in the system. This phenomenon is particularly true for a system that uses the epidemiological compartmental model as its forecasting method while using lagged data.

Publication
Working paper

Related