Forecasting Patient Demand for Improved Resource Allocation Raymond

J. Higbea, PhD & Jeffery Skinner, MHA, RN

Over several decades, demand for emergency department (ED) services has increased. Concomitantly, the number of operational EDs has decreased (AHA, 2016). As a result, this has created a resource challenge for administrations as they seek to meet increased demand with limited assets. Multiple solutions have been employed, such as redirecting less acutely ill patients to lower levels of care and providing clinical areas for patients waiting for inpatient beds (boarding) or prolonged assessments (clinical decision units). Several process improvement solutions have been employed to enhance key performance indicators, such as monitoring arrival to provider times and decision to disposition times. Other methods rely on tools to document how busy the ED is at any given time. Although these tools are accurate, they are not predictive and provide little value in resource allocation and adjustment. Finally, several EDs have found some utility in staffing to the day and shift to better meet historic resource needs; however, even these imprecise attempts fall short of expectations. The concept of value-based care began in the early 1990s as large employers, insurance companies, and the federal government attempted to rectify an imbalance between escalating healthcare costs and poor-quality outcomes as documented in reports such as To Err is Human by the Institute of Medicine (IOM, 1999). Providers were generally receptive to quality initiatives that allowed them to earn an extra 1% to 3% from payers and quality organizational certifications. Prior to these efforts, providers had begun working on improving process efficiency and patient quality. As the process and quality improvement activities matured over the next two decades, they proved themselves reckless and ineffective at improving quality or decreasing costs. With the implementation of the PPACA, the federal government began a process of paying providers for quality outcomes on a revenue-neutral basis (value-based payments)if you improved from your established base, you could earn more; if you did not improve, you earned less. As the manager of Continuum Healths metro emergency department in Majestic Falls, Michigana 100,000 plus visit per year EDpondered the current milieu, he wondered how he could better align resources with variable needs (patient volumes), while at the same time creating an ED visit of increased value (value = quality / cost) (Skinner & Higbea, 2015). The manager identified the need for variable staffing schedules that better aligned staffing resources with fluctuating patient demand. Data analysis revealed variation in patient arrival patterns by month, leading to natural 3-month groupings, creating seasonal patterns. Once these seasons were identified, they were divided by the day of the week to create four groupings of similar volumes. An existing labor forecasting tool was enhanced to create an anticipated forecast of patient arrivals, more closely aligned with typical patient arrival patterns. The original 4-hour forecast was enhanced to a 2-hour average forecast that provided the best result. Creation of a variable staffing model and the use of an enhanced labor forecasting tool allowed charge nurses to adjust staffing by 2-hour intervals, which provided significant improvements to operational effectiveness and efficiencies including:

• confidence in the planning and implementation of resources
• an annual 55% reduction in overtime pay (approximately $132,000)
• decreased quality scores variability
• decreased staff turnover
• stabilized patient experience scores (Higbea et al., 2018)

Discussion Questions

1. Based on the above scenario, explain how the deployment of this forecasting approach affected the value of an ED visit.
2. If the above forecasting model provides an ED visit of increased value, how does this increased value correlate with health policy?
3. How could policymakers use this example as encouragement for providers to seek similar management approaches?