INTRODUCTION

TO
HEALTHCARE
Shima Soltanzadeh
Session 8
OPERATING
ROOM SCHEDULING
OPERATING THEATER (OT)
The most critical cost & revenue of hospital
◦ More than 60% of hospital admissions are surgical operations
◦ More than 40% of the revenue of hospital
OR scheduling problems classification

Decision Levels
• Strategic, Tactical, Operational

Scheduling strategies
• block strategy, open strategy, modified block strategy

Patient Char.
• elective or non-elective

Problem features
• Uncertainty (duration uncertainty, arrival uncertainty, resource uncertainty
and care requirement uncertainty), objective functions, certain requirements
DECISION LEVELS

Strategic

Tactical

Operational
DECISION LEVELS
1. Capacity planning decisions
◦ number of OTs,
◦ amount of personnel,
◦ instruments (e.g. X-rays), …
2. Case mix planning
◦ assigning the OR time blocks among the surgical specialties in a long-term
time horizon
3. The dimensioning of subsequent departments’ resources (e.g. ward beds)
- Typically based on historical data and/or forecasts
- Horizon is typically long-term, e.g. a year or more
TACTICAL DECISIONS IN OPERATING THEATRES
- Input
◦ The actual aggregate patient demand (e.g. waiting lists, appointment requests for surgery)
- Output:
◦ Dividing the weekly OT time over specialties or surgeons
◦ Closed planning approach
◦ Open planning approach
- Master Surgery Scheduling Problem (MSSP) (In case of closed approach)
◦ a cyclic schedule, giving the open time of available ORs to surgeons or specialties. It
allocates OR time to surgical specialties according to their specific requirements and penalizes
undersupply.
- In a medium term, typically several weeks
TACTICAL DECISIONS - SAMPLE MSS
OPERATIONAL LEVEL
- Involving Surgery Scheduling Problem (SSP) or patient scheduling problem
- Surgeries in the waiting list are scheduled to specific OR, day and starting time.
Many studies decompose the process of OR scheduling and planning into two
steps
◦ Advance scheduling (intervention assignment or surgical case assignment)
◦ assigning an OR and a day to each surgery
◦ Allocation scheduling (intervention scheduling or surgical case scheduling)
◦ determining the starting time of the procedure
- Short-term decision making
TASKS
NURSES PHYSICIANS
- OR nurses - Each surgeon
◦ assign each room to a specialty or to an ◦ decides which cases he wishes to perform
individual surgeon. - Department heads
- OR managers ◦ confirm that the schedules proposed by
◦ responsible for managing operating their surgeons are realistic, and send them to
rooms the OR manager
- OR manager
◦ makes a final assessment of feasibility of the
proposed schedule
STRATEGIES - BLOCK SCHEDULING STRATEGY
- OR capacity is divided into slots and blocks with each OR for a specified duration
- In practice, the block scheduling is applied more often than the open strategy in hospitals,
specially in Europe.

Advantages Drawbacks
- The surgeons prefer centralizing the cases, - Other surgeons cannot occupy the block even if
◦ it is easier to make the schedule since the that surgeon doesn’t arrange any surgical cases in
work time is fixed for each surgeon the block time.
OPEN SCHEDULING STRATEGY
- Surgeons can choose to operate a case on any workday in any available OR, and no
surgeons have the priority to reserve any block time in advance.
PATIENT CLASSIFICATION
PATIENT TYPES

- Inpatients - Elective patients

◦ patients whose surgeries will be done over the
◦The patients who have to stay overnight in foreseeable future (typically 1 week) and can be planned
hospital in advance

- Non-elective patients
- Outpatients
◦ Their arrival is unexpected and should be treated as
◦Usually enter and leave the hospital on the soon as possible

same day
INPATIENTS AND OUTPATIENTS
- Inpatients are admitted to the hospital one or more days before surgery and
stay in the hospital after surgery for continuing care,
- Outpatients usually have same-day surgery and therefore do not stay in
hospital overnight
- Difference in scheduling
◦ The inpatients can be considered as stand-by while the outpatients
dynamically arrive at the hospital with possible lateness, no show or cancel
◦ The inpatients satisfaction is closely related to the schedule operation date,
while outpatients have lower requirements on it
ELECTIVE AND NON-ELECTIVE
- Many researchers do not point out explicitly what type of patients they mainly target at
- Most papers focus on elective patients
- Two approaches of reserving OR time for emergency surgery
◦ Reserving the dedicated emergency ORs and an emergency case is operated immediately
if the dedicated OR is empty
◦ Emergency patient is performed once one of the ongoing elective surgeries has ended
- A discrete event simulation
◦ The results show that emergency patients are operated upon more efficiently on elective
Ors instead of a dedicated Emergency OR.
PROBLEM FEATURES-UNCERTAINTY

1. Duration uncertainty

2. Arrival uncertainty

3. Resource uncertainty

4. Care requirement uncertainty

UNCERTAINTY
1. Duration uncertainty
◦ Deviations between the actual and planned durations of relevant activities
◦ By the patient condition, and the skill of the surgeon, surgical specialty such as orthopedic, …
◦ Objective: minimizing the overtime and idle time of the ORs
◦ Three distributions to model the surgery duration: the lognormal, gamma, and normal

2. Arrival uncertainty
◦ The unpredictable arrival of outpatients at the start of the surgery operation

3. Resource uncertainty
◦ A set of resource combination is open to certain cases on certain time in certain place, while others may be
unavailable or inapplicable.

4. Care requirement uncertainty

◦ Assigning patients with the objective of minimizing excess workload for nurses
OBJECTIVE FUNCTIONS
OBJECTIVE FUNCTIONS
MAXIMIZING EFFICIENT USAGE OF THE MINIMIZING THE SUM OF THE EXPECTED

OPERATING THEATRE WAITING COSTS FOR PATIENTS

- Expected idling cost of the resources - Minimize patient-related costs
- Maximize the use of ORs time - Minimize the costs of patients waiting time
- Minimize the cost of ORs - Maximize the patient satisfaction
- Minimize the number of ORs
- Minimize the maximum makespan
- Minimize the overtime
- Minimize the number of used blocks
- Minimize the transitions of surgeons between different ORs
AN EXAMPLE-PHASES
- Phase1: Session Planning Problem (SPP)

◦ Available OT time is distributed among the wards(specialties)

- Phase 2: Master Surgical Schedule (MSS)

◦ A cyclic timetable that determines the surgical unit associated with each block of OT time

- Phase 3: Elective Case Scheduling (ECS)

◦ Sequencing and scheduling elective cases in each block

ASSUMPTIONS
1. Fixed number of surgery rooms

2. Fixed number and length of OT blocks available for the elective surgery

3. Only one ward should be assigned to a surgery room

4. Emergency patients and outpatients are not considered

5. Uncertainty in demand as well as length of stay duration only in operational level

SESSIO N PLANNING PRO BLEM (SPP)
- Determining the number of sessions to be scheduled weekly for each ward

Like which well-known model?

packing problems

◦ Bin-packing problem

◦ (Multi-) Knapsack problem

BIN PACKING PROBLEM
- Given n items with sizes s1, s2, ..., sn such that 0 ≤ si ≤ 1 for i ≤ i ≤ n, pack them into the

fewest number of unit capacity bins.

EXAMPLE APPLICATIONS
- Given n items with sizes s1, s2, ..., sn such that 0 ≤ si ≤ 1 for i ≤ i ≤ n, pack them into the

fewest number of unit capacity bins.

BIN PACKING MODEL
Problem is NP-hard (NP-Complete for the decision version).
MULTIPLE-KNAPSACK PROBLEM
- Given n items, with each item j having an associated profit pj ,weight wj. The binary
decision variable xij is used to select the item j for knapsack i.
- The objective is to pick some of the items, with maximal total profit, while obeying that the
maximum total weight of the chosen items must not exceed ci.
MULTIPLE-KNAPSACK VS. BIN PACKING
MULTIPLE KNAPSACK BIN PACKING
- Knapsacks are fixed - Bins are variables
- Maximizing the profit of packed items - All items are packed
SPP MODEL
Maximizing the total benefit achieved by the OT supply
◦ Ward demand Dj
◦ Marginal benefit djk of the kth session
demanded by the jth ward
SPP MODEL
TWO-STEP PROCESS OF BENEFIT
1. Compute the ward demand based on historical data and waiting list,

DSj: number of sessions which have been assigned weekly to ward j ward j in the past,
DLj: the number of sessions necessary to clear the waiting list of ward j.
2. Compute the “marginal” benefit djk
◦ kth session demanded by the jth ward
PHASE 2: DESIGNING A MASTER SURGICAL SCHEDULE
- Which surgical specialties operate on which days?
- Input: Nj of the first phase, number of sessions for ward j
- A MSS represents a repetitive pattern over a certain number of days (a cyclical assignment
of OT time)
PHASE 2: DESIGNING A MASTER SURGICAL SCHEDULE
- Objective function
◦ maximizing the surgeon preference
◦ Pjw : preference of surgeon in ward j for day w
How to measure Pjw ?
- Surgeon utility function depends on:
◦ exclusion of particular days due to other engagements
◦ expected patient length of stay (short or long stay)
◦ WLssj: Number of short stay patients on the waiting list of ward j
MSS FORMULATION
Assignment constraint: each room only to one ward during the day
Assignment constraint: restricting number of rooms in each ward to available teams on day
Capacity constraint: restricting OT sessions of each ward to the number of weekly sessions
selected in phase 1
OUTPUT
PHASE 3: ELECTIVE CASE SCHEDULING (ECS)
- How to schedule patients into the assigned OT sessions?
Approaches
◦ Mathematical model
◦ Distribution rules
◦ Longest waiting time (LWT)
◦ Longest processing time (LPT)
◦ Shortest processing time (SPT)
THREE SCHEDULING RULES
◦ Shortest processing time (SPT)

◦
Longest processing time (LPT)

Longest waiting time (LWT)

DISCRETE-EVENT SIMULATION MODEL
PHASE 3: ELECTIVE CASE SCHEDULING (ECS)
- Patients as entities - Performance index
◦ Clinical status ◦ number of operations performed in time T
◦ Expected Operating Time ◦ hours of overrun
◦ Expected Length of Stay ◦ number of shifted operations
◦ Updated waiting time ◦ bed utilization rates
- MSS mode ◦ short stay, ρSS(%)
◦ Mode 1: Actual hospital MSS ◦ long stay, ρLS(%)
◦ Mode 2: From phase 2
SIMULATIONS RESULTS