June 2018
Spotlight: Workforce Optimization
ic_arrow_up.svg
Hide Page Sections
ic_arrow_drop_down_black_24px.svg
Show Page Sections
While eliminating all nursing workforce overtime spend is extremely difficult, the Workforce Optimization approach minimizes such spending. It also helps organizations reduce contract labor costs by creating an in-house float pool that can be used to flex nursing staff for periodic spikes in patient demand, a portion of non-productive coverage, vacancies recognizing time-to-fill, demand seasonality, and unscheduled absences. This methodology can also be applied to systems looking for an enterprise-wide staffing option. For example, by modeling all medical surgical units across business units, a regional float pool can be right sized and configured.
Model 3: Nursing Workforce Schedule

Next, Model 3 produces a de-identified scheduling pattern for the hospital’s unit-based nursing workforce. This provides the “optimal schedule” for each individual nurse, and adheres to the hospital’s labor work rules (i.e., nurses are required to work every third weekend, nurses cannot work three consecutive days, etc.). This model uses an algorithm to compare thousands of possible scheduling options and identify the best solution.
Model 2: Nursing Workforce Supply

Following Model 1, Model 2 is used to produce the optimal number of FTEs. Optimization methods are applied to determine appropriate levels of unit-based staff and float pool staff for the hospital. Float pool nurses float from unit to unit to aid in patient census surges, rather than relying on more costly contract employees.
Model 1: Patient Demand

This model involves taking hourly patient census data for each hospital unit, and producing forecast models for each unit. The forecast models should utilize Maximum Likelihood Estimators, and go beyond the “average” used in the traditional approach, taking into account fluctuations and seasonality in patient demand. Each unit produces multiple forecast models using a single composite method of Tabu Search, Neural Networks, Scatter Search, and Linear and Integer Programming. The “best” model for a unit is selected based on a series of statistical measures.
Any nursing manager will tell you creating a schedule for his/her unit is a tall order each week. They spend countless hours compiling a schedule that ensures each nurse meets his/her FTE requirement, complies with the hospital’s work standards (i.e., working every third weekend, working no more than three consecutive shifts, etc.), and provides adequate nursing coverage to meet the budgeted ADC. Any holes in the schedule typically are filled with premium labor in the form of contract nurses, overtime shifts, and/or premium pay shifts. Creating such schedules is a menial task that takes valuable time away that managers could be focused on patients. This article describes a more effective approach to the nursing workforce problem, which considers more than just the ADC taken at midnight, HPPD, and a hospital unit’s nurse-to-patient ratio. It utilizes advanced analytics and statistical modeling to produce a more flexible and affordable workforce. The Workforce Optimization approach incorporates a vast amount of data, including a hospital’s hourly patient census, hourly timekeeping, payroll, budget, staffing grids, position control, and employee data for the past three calendar years. Qualitative data also is incorporated to better understand the hospital’s staffing parameters (“work rules”). This model considers many of the working parameters needed to take a mathematical model from concept to reality, and aims to automate all of the manual parameters that have to be considered when constructing a schedule. The model is comprised of more than 50 inputs per hospital unit, such as the unit’s historic census, bed capacity, nurse-to-patient ratio, operating hours, non-productive rate, vacancy rate, FMLA rate, weekend working requirements, and weekend start and end times. Once these inputs have been finalized for each unit, a three-step modeling process is used to better address the nursing workforce problem:
Workforce Optimization
Case Study
The Nursing Workforce Optimization model was applied at a multi-hospital system in the Chicago metropolitan area in 2015. The project encompassed roughly 13,000 nurses and three regions where float pool staff would travel. The first step was to standardize work rules across the system. Representatives from each hospital came together and established system-wide rules, such as requiring nurses to work every other weekend, allowing for a mix of 12-hour and 8-hour staff, and defining the skills needed to produce a high-performing float pool. Last year, 25 percent of all nursing hires were for float positions, moving the organization closer to the goal of creating a more flexible clinical workforce. To help enhance and recruit for the float pool, the organization’s leaders changed the image of the float pool from “excess staff” to “elite staff.” A highly skilled team also was brought in to help remedy extreme circumstances. As a result of Workforce Optimization approach, the health system reduced overstaffing and understaffing by 22 percent the first year, and saved $16 million in nursing labor spend (equivalent to more than 5 percent of its nursing budget), including a $9 million reduction in premium labor. The system’s savings came from having FTEs better aligned to reduce spending; only one FTE was eliminated as a result of the outputs. In addition to the positive financial results, associate engagement increased from the 39th to the 54th percentile, and units meeting their HPPD targets increased from 50 percent to 80 percent.
A macroeconomic view of Contract Hours as a Percent of Paid Hours shows an upward trend over the same time period. Many hospitals have shifted to creating a more flexible nursing workforce, but they are attempting to do this with contract labor—a very expensive resource pool—rather than creating a flexible pool of their employed nurses. It is also important to note the importance of determining the root cause of premium labor utilization. Frequent drivers include insufficient non-productive coverage, poorly or inadequately configured float pool coverage, temporary holding of budgeted positions or unrecognized demand seasonality.
A Look at the Data

A macroeconomic view of Overtime Hours as a Percent of Productive Hours shows that trends have remained relatively stable from January 2017 to May 2018, with spikes in January and slight dips in April both years. This indicates that budgeting for nursing and variable staff remained largely unchanged, and an inflexible workforce still exists.
This approach creates a nursing workforce that is not very flexible, and in turn, generates potential for over- and under-staffing, which are detrimental to both patient and staff satisfaction scores. An inflexible workforce leads to high labor costs, causing many organizations to spend more than they have budgeted. These data also point to a frequent phenomenon, often present with 12 hour schedules, whereby a unit may be staffed adequately on average, however over staffing is present earlier in the day as demand slowly builds through the course of a day. By late afternoon the unit may be understaffed suggesting a need for atypical shift lengths or start times.
Though there is marginal productivity improvements in May, the heating up of the labor market potentially poses challenges to hospitals. Most hospitals and health systems take a traditional approach to budgeting and composing the nursing workforce. This approach relies on two key metrics: the Average Daily Census (ADC) and Hours per Patient Day (HPPD). The ADC—the number of patients on the nursing units—typically is taken at midnight and applied as the “average” for an entire work day. This measure then is lined up against a nursing unit’s target HPPD to determine the number of FTEs the unit needs to provide adequate patient coverage.

Conceptually, it is easy to see where the logic of this approach is flawed. Applying the patient census at one hour of the day as an average does not account for the other 23 hours in the day. It doesn’t account for fluctuations in the patient census that may impact staffing, nor does it capture the potential seasonality or day of week variation on the nursing unit.
Background
chart_1_v2
Overtime Hours as a Percentage of Productive Hours | January 2017 to May 2018
Contract Hours as a Percentage of Paid Hours | January 2017 to May 2018
Workforce Optimization Models
Case Study
Background
mail_gray.svg
Share article
Workforce Optimization Models
Case Study
Background
mail_gray.svg
Share article
mail_gray.svg
Share article
©2018 Kaufman, Hall & Associates, LLC
Become a subscriber
Sign up now for access to the latest news and reports
Become a subscriber
Submit
Annual subscription
Enterprise subscription
Receive our monthly reports
Your details
kha_logo.svg
National Hospital Flash Report
menu_icon.svg
Become a subscriber
Submit
Annual subscription
Enterprise subscription
Receive our monthly reports
Your details
scroll_up.svg
Sign up now for access to the latest news and reports
SUBMIT
Become a subscriber and receive our monthly reports. Choose the option that best fits your needs.
SUBMIT
Monthly subscription: $100/month
Annual subscription: $600/year
Quarterly subscription: $200/quarter