STAFF SCHEDULING

Serhat Gul, Ph.D.

TED University
IE 482 Decision Making in Health Care

Main references:
Burke, E., P.D. Causmaecker, G.V. Berghe, H.V. Landeghem. 2004. The state of the art of nurse rostering.
Journal of Scheduling.7.441-499.

Ozcan,Y. Quantitative Methods in Health Care Management. Chapter 8 1

Remember the three main steps in
workload management!

• Staffing-- determining the appropriate number of full-

time equivalents (FTEs) to be hired in each skill class (RN,
LPN, aides, MHA, MBA, etc..)
• Scheduling-- who is on and off duty and when
• Reallocation-- fine tunes the previous decisions as
uncertainties resolve at each shift or day

2
 STAFF SCHEDULING
• Staff Scheduling involves the allocation of the budgeted FTEs
to the proper patients in the proper units at the proper time.
– Coverage
– Schedule Quality
– Stability
– Flexibility
– Cost

3
 Scheduling Types
✓Cyclical work schedules employees do not rotate
shifts planned for 4-6 week period; repetition provides
stability and lower scheduling costs, but is inflexible
best in a stable environment

✓Discretionary/flexible work systems

➢staggered start-- does not change # of hours
worked; employee chooses when she/he starts
➢staggered week-- average 40 hrs a week 8 hr
days, but alternate weekly schedules (e.g.., 4/32
and 6/48)

4
Exhibit 8.1 Cyclical Staffing Schedules for Four Weeks

The Four Week Schedule

Employee S M T W T F S S M T W T F S S M T W T F S S M T W T F S
A 0 0 0 0 0 0 0 0

B 0 0 0 0 0 0 0 0

C 0 0 0 0 0 0 0 0

D 0 0 0 0 0 0 0 0

# of Staff 2 3 3 3 3 3 3 2 3 3 3 3 3 3 2 3 3 3 3 3 3 2 3 3 3 3 3 3
Scheduled

5
Flexible Scheduling Systems

Very common in health care

Part-time shifts and float pools are

necessary to meet staffing needs

Enhanced by computerized
scheduling technology

6
How long should FTEs work?
Advantages of various scheduling patterns.
8 10 12
Traditional X
Several consecutive days off X X
More weekends off X X
Increased personnel during X
busy periods
Staff able to take adv. of X
meetings, continuing educ.
Increase in care continuity X
Fewer staff positions needed X
Recruitment drawing card X

7
How long should FTEs work?
Disadvantages of the various shift patterns.

8 10 12
Fewer weekends off X
Fewer consecutive days off X
Requires increased staff X
Possibility of increased X X
fatigue times
Appearance of less time X X
for relaxation
Administrators plan for X X
3 shifts

8
Figure 8.1: Comparison of 8 and 10 Hour Shifts

Overlapping
Nights Evenings Shifts Days

10 Hr.
Shifts

8 Hr.
Shifts

7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7

9
Figure 8.2: Pattern of Alternating Eight and Twelve Hour Shifts

Days Evenings Nights

Mon Tue Wed Thu Fri Sat Sun

10
 Nurse Rostering
• Definition: The allocation of nurses to periods of work over
several weeks (or short-term timetabling of nurses)
• Typical planning period: 4 weeks
• Commonly used shift type: (7am-3pm), (3pm-10pm), (10pm-
7am)
• Coverage constraints: Number of personnel needed for every
skill category and for every shift or time interval during the
planning period
• Time related constraints: All restrictions on personal
schedules: personal requests, personal preferences, workload
balance among personnel
• Work regulations: The contract that personnel members have
with the hospital
 Nurse Rostering Literature

• Papers on computerized health care personnel scheduling

problems have been published since 1960’s
• Early works developed deterministic models for simplified
problems (Miller, Pierskalla and Rath 1976, Trivedi and Warner 1976, Warner
1976, Warner and Prawda 1972)
• Several heuristics have been developed for solving real world
problems (Anzai and Miura 1987, Isken and Hancock 1991)
• Artificial intelligence techniques have been used since 1980’s
(Chan and Weil 2001, Chiarandini, Schaerf and Tiozzo 2000)
Methods for Nurse Scheduling Problem

• Optimization methods:
– Linear and integer programming
– Goal programming / multi-criteria approaches
• Finding optimal solutions is infeasible and/or meaningless
• Administrators want to quickly generate a high quality
schedule that satisfies all hard constraints and as many soft
constraints as possible
• Alternative methods:
– Expert systems/knowledge based systems
– Heuristics/Metaheuristics
Nurse Rostering (Example 1)
• In order for you to gain full understanding of the problem (just for
illustrative purposes), one particular instance for the problem is defined as
follows: There are ten nurses working in an outpatient clinic (n=10) and
the total regular required working time per nurse over the planning
period of one month is 160 hours (RWTi = 160) for all nurses (i = 1,…,10).
There are 30 days in the month we consider as the planning period. The
first day of the month (when j=1) is Monday. There are 7 different shift
types used in this outpatient clinic. Each day is divided into 6 demand
periods and the demand varies across periods. The shift and demand data
is provided in the below table.

15
 Nurse Rostering Example

Chapter 7: Quantitatve
Methods in Health Care Yasar A. Ozcan 16
Management
• There are some aspects of nurse rostering that are not
considered in the model.
– Extension 1: For example, this point is important due to work
regulations: The total work time per any consecutive 7 days
(corresponding to a week) is at most 50 hours for each nurse. Revise
the above model based on this additional hard constraint.
– Extension 2: Suppose that the shift H1 represents the first shift on a
day, while the shift HS represents the last shift on a day. Then,
formulate a linear constraint to consider the following restriction: If a
nurse works the last shift on a day (for example, day j ), then she can
not work the first shift on the following day (i.e. day j+1).

Chapter 7: Quantitatve
Methods in Health Care Yasar A. Ozcan 17
Management
 Should the staffing and rostering
decisions be coordinated?
• Theoretically yes, but it is impractical:
– Even though there is a high fluctuation in patient needs, it is not
recommended to shift personnel around the hospital each time the
request does not match the available time. This would be the
consequence if the problem were looked at from a purely global point
of view.

– People prefer to express personal preferences with respect to work

and free time. These preferences differ from time to time.

– Trade-off between finding an optimal solution in a large amount of

time vs. suboptimal solution in a reasonable amount of time
 Nurse Rostering (Example 2)
• Four different skill levels:
– 1 head nurse, 15 regular nurses, 3 caretakers, 2 trainees
• Shift types:
– Early, day, late, night
• Work regulations:
– Full-time nurses: 38 hrs/week; max 6 assignments to night shift, 2
assignments to weekends per month.
– Part-time nurses: 20 hrs/week; max 10 assignments per month
– A nurse trainee should be in the same shift as her supervisor
• Nurse preferences:
– Request for a particular day off

Burke, E., P.D. Causmaecker, G.V. Berghe, H.V. Landeghem. 2004. The state of the art of nurse rostering.
Journal of Scheduling.7.441-499.
 Nurse Rostering (Example 2)
• Planning period:
– 4 weeks
– Head nurse needs to create a roster
• The coverage constraints are imposed by the management
– Should be met for each skill category and shift type
• Quality aspects:
– Assignment to consecutive shifts
• Objective:
– Minimize the level of violations of time related constraints
 Administrative modes of
scheduling process
• Centralized scheduling:
– One administrative department in a hospital carries out all the
personnel scheduling
• Unit scheduling:
– Head nurses or unit managers generates the schedules locally
• Self-scheduling (i.e. interactive scheduling):
– Personnel roster is generated manually by the staff themselves
Ronnberg, E. and T. Larsson. 2010. Automating the self-scheduling process of nurses in
Swedish healthcare: a pilot study. Health Care Management Science. 13. 35-53

Self-scheduling process
Mathematical model for self-scheduling
 Important Notes

• Automatic generation of nurse schedules can lead to

improvements in resource efficiency, staff and patient safety,
staff and patient satisfaction, and administrative workload.
• Many of the models are too simple to be directly applied to
hospital wards.
• The gap between the capability of the models in the literature
and demand of administrators must be handled.
An interesting paragraph!
 References

Ronnberg, E. and T. Larsson. 2010. Automating the self-

scheduling process of nurses in Swedish healthcare: a pilot
study. Health Care Management Science. 13. 35-53

Burke, E., P.D. Causmaecker, G.V. Berghe, H.V.

Landeghem. 2004. The state of the art of nurse rostering.
Journal of Scheduling.7.441-499.

Warner, D.P., J.Prawda.1972. A Mathematical Programming Model for

Scheduling Nursing Personnel in a Hospital. Management Science. 19
(4). 411-422