The COVID-19 outbreak has magnified some key scheduling challenges in healthcare delivery systems such as overcrowded ERs, long patient wait times, overworked frontline staff, and hidden capacity. Scheduling challenges will continue in the post-pandemic world as providers rush to reschedule previously cancelled procedures. These challenges are common for all types of health services. In units serving walk-in patients, such as ERs, it is a struggle to forecast volume and serve demand while addressing individual staff needs. Scheduling appointments can be just as difficult. It only takes a few no-shows or late arrivals to throw a wrench in a planned schedule. The silver lining is that scheduling is not unique to the healthcare industry, and there is no need to build solutions from scratch in this digital age.
Inspiration from the retail industry
Fluctuating customer demand, late or no-show customers, and complicated shift schedules are challenges faced by retail operators from cinemas to stores to restaurants. The retail industry has improved forecast and optimization capabilities through machine learning, automation, and other digital solutions. A Big Box retailer drastically improved their demand forecasting by layering weather data with individual store demand to their benchmark after observing a decrease in foot traffic on rainy days, enabling optimized shift schedules and improved e-commerce operations. Large retail stores like Walmart and BestBuy are using staff self-scheduling applications to streamline scheduling and employee communication. These lessons learned and tools used by the retail industry can spark innovation and be adapted for the healthcare industry.
Five key capabilities to address healthcare scheduling challenges
Despite the commonality, each health service delivery unit has its unique challenges across patient and staff, from scheduling planning to daily schedule operations. We identified five key capabilities that address these challenges – each with varying need for healthcare specific sophistication.
Accurate demand forecasting lays the foundation for schedule optimization. Important attributes to look for when evaluating solutions are the level of granularity for patient inflow, models that can place more weight on recent data to capture changes such as a new competitor nearby, ability to run various simulations based on inputs such as acuity, treatment needs, and ability to incorporate evolving targets and metrics. The technology is widely available – the challenge lies in aggregating accurate internal, external, structure and unstructured data and building the machine learning models to spot complex relationship between the datasets for each unique operation.