Creating Specific Plans in an Uncertain Environment: An Example

Here is an example of what one medical center did to create a staffing strategy for patient care staff in the face of great uncertainty. Note that the approach utilizes several of the techniques just described, not just one. The medical center needed to define staffing requirements (and develop staffing strategies and plans) even though its patient load varied significantly from day to day. It was relatively easy for the center to calculate a staffing ratio that specified the number of patient care staff (e.g., registered nurses) required per patient in a given unit. That was not the problem (some states even mandate or regulate these ratios). The problem was that the center had very little idea of how many patients could be expected in any given unit at any given time. While the number of patients in each unit was not random, it fluctuated greatly on a daily basis. The center had attempted to predict the number of patients in each unit on given days (e.g., forecasting based on cyclicality and seasonality), but these efforts had proved fruitless. Consequently, staff planning seemed impossible. There was no clear way to determine a daily patient census to which the known staffing ratio could be applied. Schematically, the situation in one unit looked like the graph in Figure 8-1.

This graph indicates that the number of patients fluctuated significantly over the course of a year, following no particular pattern or cycle. At first, the daily patient census seems to be virtually random. However, the medical center realized that no matter how random the number of patients seemed, there were at least three pieces of critical information that were known for sure. First, historically the number of patients in this unit had never been less than 60. Similarly, the number had never exceeded 100. Finally, the overall actual mode (i.e., the number of patients most often observed) was 86 patients.

Suppose we were analyzing staffing requirements for nurses. As stated previously, the medical center had already established specific staffing ratios for each patient care staff position. If we assume (for example) a 2:1 patient/nurse ratio, we can determine that the minimum number of full-time equivalents (FTEs) required for this unit is 30 (i.e., 60 patients/2 patients per nurse), the maximum is 50 (i.e., 100/2), and the mode is 43 (i.e., 86/2). What staffing level was appropriate?

One alternative was to calculate and staff to the maximum, but this would be inefficient, and the cost would be prohibitive. Employing this alternative would create staffing surpluses (some of which would be significant) virtually every day. Another option would be to staff to an average number of patients, but this would mean that half the time the unit would be overstaffed (creating an expense that could not be tolerated) and half the time the unit would be understaffed (with dire impact on patient care and satisfaction). A third option would be to staff to the minimum, but then what would be done on those days (in fact, virtually every day) when there were more than a minimum number of patients?

The staffing strategy that the medical center actually devised had three parts. First, each unit would be staffed with the hospital’s full-time staff up to the minimum required number (in this case, a base staffing of 30 FTEs). Second, whenever the number of patients exceeded the minimum but was less than the mode, the hospital’s own part-time staff would be used. This was the level of uncertainty the hospital was willing to accept and deal with. For the unit in the example, that meant that up to 13 FTE nurses would be required from the center’s part-time pool on any given day, where 13 equals the difference in staff needed to support the mode (43 FTEs) and the minimum (30 FTEs). Third, any staffing above the mode would be obtained from an external registry. In this case, required staffing levels could range from 0 to 7 FTEs, where 7 equals the difference in staff needed to support the maximum (50 FTEs) and the mode (43 FTEs). In effect, the hospital decided to outsource the greatest uncertainty regarding required staffing to its partner, the staffing agency.

This approach allowed the hospital to create some very specific staffing plans, even though there was great uncertainty regarding the number of patients that were expected on any given day. Because full-time staffing was based on a minimum that did not change, staffing plans for full-time staff could be developed completely, and in great detail. Given the approach that was taken, 30 FTEs were required every day, every shift—no matter how many patients there were.

There was some uncertainty regarding the number of part-time staff needed, but this approach allowed the hospital to define the size of the part-time pool that would be needed to implement this strategy. Once it knew the range for the number of part-time FTEs that would be needed (i.e., none when patient count was at the minimum up to a maximum of 13 FTEs when patient count was at the mode or above), the hospital could scope out the size of the part-time pool that would be needed to provide the required number of FTEs each day. In fact, the hospital realized further efficiencies (and reduced uncertainty even more) by creating common pools of part-time staff for units that had similar needs (e.g., one pool for medical-surgical units, another for the cardiac care unit and the intensive care unit, and a third for the emergency department). Because it was unlikely that all the units served by a given pool would require minimum (or maximum) staff at the same time, staffing requirements tended to balance each other and the number of staff in the pool could be modulated. In effect, there would always be a need for some staff, and there would never be a need for maximum staffing in all units.

There was even more uncertainty regarding the external pool (i.e., half the time none would be needed; the rest of the time as many as 7 FTEs might be needed), but this uncertainty was, in effect, absorbed by the agency. Because the hospital could provide the staffing agency with fairly specific parameters regarding its potential need (e.g., the maximum size of the pool, the skills required, and the turnaround time expected), the agency knew what was expected and could price the contract accordingly.

In the end, the hospital was able to develop and implement a very specific staffing plan even though it had no idea how many patients to expect on any given day.