• Business Process Management
• Health Care
• Human Resources
• Call Center
• Logistics/Supply Chain
• SIMPROCESS Features
• Demonstration Models

Health Care

			Emergency Room Model Overview
			Optimizing Rehabilitation Patient Scheduling
			Architecture Analysis and Modeling for Banfield
			Point of Dispensing (POD)

  Emergency Room Model Overview

The administrators of an Emergency Room need to find the optimal staffing levels. The Emergency Room must be able to treat its patients in a timely manner, yet not be overstaffed (which costs the hospital a lot of money). Therefore, a simulation model of the Emergency Room was built in SIMPROCESS in order to predict optimal staffing levels.

The simulation model diagramed the current process ("As-Is"). Patients arrive in the Emergency Room either through the entrance door or via ambulance. The hospital groups these patients into three categories - levels 1, 2 and 3. Level 1 patients, such as heart attack victims, are considered the most critical and need to be treated immediately. Level 2 and 3 patients go through a triage process where the hospital makes an initial assessment of their injuries. All patients are then transferred to an available room for treatment. They also go through a registration process either before or after treatment, depending on the severity of their injury. After initial medical treatment the hospital can release the patient or assign them to a room in the hospital for a longer stay.

Solution

The Emergency Room needs to be adequately staffed to meet the numbers of incoming patients. Historical data were used to simulate the number of incoming patients broken out into arrivals by day, evening and morning, and the severity of their injuries. To satisfy these arrivals some of the resources have fixed levels - i.e., one charge nurse, one triage nurse and 13 wheelchairs. However, the other resources do not have fixed levels. The SIMPROCESS model was used to find the optimal levels for each given the following constraints

Administrative Clerks - Min 1, Max 4
Nurses - Min 1, Max 7
Patient Care Technicians (PCTs) - Min 1, Max 4
Physicians - Min 1, Max 6
Emergency Rooms - Min 1, Max 20

In order to find the optimal levels of these resources the SIMPROCESS optimization tool, OptQuest, was applied to the Emergency Room Model. The objective for OptQuest was to "maximize the rooms in use," which would in effect reduce the number of unneeded resources and rooms. OptQuest ran 15 iterations of the Emergency Room Model. The optimal results were

Administrative Clerks = 4
Nurses = 6
Patient Care Technicians (PCTs) = 4
Physicians = 6
Emergency Rooms = 16

The hospital applied those staffing levels to the Emergency Room model and was able to determine the associated cost. In addition, a SIMPROCESS dashboard was used to graphically illustrate the proposed staffing levels by showing the "Room Usage" and "Number of Patients Waiting." The dashboard clearly showed the Emergency Room's resources would provide timely and appropriate service for all incoming patients.

Demonstration Model - Emergency Room (EmergencyRoom.spm)

Emergency Room Overview - Robo Flash Demo

Return to Top

Optimizing Rehabilitation Patient Scheduling

Executive Summary

This example focuses on the use of process simulation for optimizing rehabilitation patient scheduling. A proof of concept scheduling model for a major health care system provider in the Southeastern U.S., Eastern Health Systems, Inc. is constructed using a simulation tool and process characteristics like patient throughput, staff productivity, labor costs, etc. are compared to assess the effectiveness of centralized versus decentralized rehabilitation patient scheduling.

Purpose

The purpose of this proof of concept paper is to report results generated by rehabilitation patient scheduling alternatives. Workflow aspects like staff productivity, patient throughout, labor costs, etc. are compared to assess the effectiveness of multiple solutions. The model developed using the SIMPROCESS simulation tool was based on data made available by health care experts. The model was built using an iterative approach and validated with the subject matter experts. Credibility of the simulation was established by comparing model results to the empirical (actual) set of data. A centralized scheduling alternative was further developed.

Goals and Objectives

The primary goal of this proof of concept effort is to use process simulation as a mechanism for experimenting with improvement ideas in the health care area. A centralized scheduling scenario was developed in this context. A second goal of this proof of concept initiative was to show that an engineered approach could potentially provide additional value to subsequent traditional project management processes.

Examples of those processes include

• Alternative options analysis for implementation strategies, including business process changes
• Identification of change management and training needs to support any changed business processes
• Product evaluation and selection for multiple vendor solutions

• Business Process Reengineering (BPR)
• Event Simulation
• Queuing Theory

In our approach to redesigning the rehabilitation outpatient workflow we used a repeatable process that is part of our BPR methodology. This is a formalized process that will help Eastern Health Systems, Inc. modernize and improve their processes/supporting information technology requirements. The SIMPROCESS business process simulation product was used to speed up the data gathering, simulate "What-If" scenarios and provide key metrics for decisions on "To-Be" process planning.

BPR Methodology

This approach is based on five phases that are executed iteratively through each business area to provide an incremental approach to BPR. The incremental approach has proven to be more risk adverse than big-bang, top-down approach to reengineering. The five phases defined are

1. Understand the Current Health Care Business Environment
   This phase focuses on setting expectations, identifying scope of the project, defining goals of the BPR effort and defining the problems to be focused on during the BPR analysis. The BPR was performed in increments to produce benefits rapidly and to reduce risk by breaking the process into more manageable pieces.
   
   
2. Model the Current Health Care Processes ("As-Is")
   This phase focuses on the analysis of the existing legacy health care processes and in developing the "As-Is" models. The legacy processes, organizational structure and roles of the organizations were documented, measured and baselined for comparison to the "What-If" models (centralized rehabilitation patient scheduling). Metrics required to support development of the "As-Is" rehabilitation patient scheduling model were collected through a series of interviews and workshop sessions with health care experts.

Items that were captured using the SIMPROCESS simulation tool include

• Process flow
• Frequency of actions
• Level of effort for each action
• Resources performing each action
• Source of information for each action
• Cost information

The first cut "As-Is" rehabilitation patient scheduling model was validated and verified by the process owners of Eastern Health Systems, Inc. Validation involved review and consensus that the process flow, resources and level of effort were acceptable. Once the process flow was validated then the results of the simulation (quantities, task times and staff productivity) were verified with the knowledge experts of Eastern Health Systems, Inc. The results of the "As-Is" rehabilitation patient scheduling model provided a quantitative baseline against which alternatives were evaluated.

3. Visualize and Measure Alternatives ("What-Ifs")
   This phase focuses on the analysis of potential BPR improvements that were considered for implementation. Improvements typically include organizational changes, resource assignments, roles and responsibilities, process flow changes, insertion of technologies, improved access to information and changes to policies/procedures. The "What-If" rehabilitation patient scheduling model was specifically developed to experiment with changes in the scheduling process. The centralized alternative model metrics were captured and compared to the baseline for potential quantitative benefits such as savings in cost and patient throughput. The most important metric produced from the simulations was Activity-Based Cost (ABC). The ABC metric is a key ingredient in the development of a ROI for the "What-If" rehabilitation patient scheduling model and in justifying the development of the "To-Be" plans for implementation.
   
   
4. Plan the Transition ("To-Be")
   This phase focuses on choosing the best value scenarios from the "What-If" modeling and developing implementation plans for them. The metrics from the "What-If" model will provide the basis for the choices and will assist Eastern Health Systems, Inc. in deciding which alternative should be implemented to get the most ROI. Other considerations are also factored into the decision process. Such factors as degree of reorganization, cost of training, cost of process implementation, time to implement, cost of new technology insertion, etc. are examples of additional input to the decision process. For example, a "What-If" scenario that clearly provides huge ROI benefits may be too risky to be implemented first due to the organizational or cultural impact. In this example, a lesser ROI-based "To-Be" implementation may be planned as the first increment to avoid risk of cultural change and revisited later in the implementation stages.
   
   
5. Manage Transition and Monitor Performance
   This phase focuses on the actual implementation of the improvements. The simulation model will be used to assist in the identification of development phases and for planning the transition. The SIMPROCESS simulation tool will be used to ensure that the process is not broken as changes are fielded and implemented.

Model Overview

Top Level Processes - Figure 2 shows the top level view of the health care simulation model. The model uses labels to display the number of

1. "Total Rehabilitation Patients"
2. "Total Rehabilitation Visits," and
3. "Average Rehabilitation Visits/Patient"

Icons representing entities (i.e., patients) flow through the sub-processes where delays capture the amount of work time required for resources (i.e., clerks, doctors, etc.) to perform a given task. For this proof of concept effort the processes considered core were Rehabilitation Scheduling and Rehabilitation Therapy.

Top Level Processes

"Rehabilitation Scheduling" Process Box Icon Details
The rehabilitation scheduling process box icon contains the Scheduling and Registration sub-processes.

Details of "Rehabilitation Scheduling" Process Box Icon

"As-Is" Rehabilitation Patient Scheduling Alternative (Decentralized)
This figure represents the details of the legacy scheduling process alternative (Scheduling Process Box Icon-Alt1).

"As-Is" Rehabilitation Patient Scheduling Alternative

"What-If" Rehabilitation Patient Scheduling Alternative (Centralized)
This figure represents the details of the centralized scheduling process alternative (Scheduling Process Icon-Alt2). The "pre-registration" step is removed and included as part of the "collect patient data" activity.

"What-If" Rehabilitation Patient Scheduling Alternative

Model Settings

This proof of concept model was set to run for a 10-day period for one replication. Costs were reported at the end of the simulation run. Statistics were captured for the patient arrivals, patients in process and patients completed.

Conclusion

The health care simulation model provided insight into the process map and allowed for the subject matter experts to refine the flow based on the results of the baseline. Model metrics were collected through a set of interviews and workshop sessions and performance characteristics such as rehabilitation patient turn-around time and labor costs were discussed with the knowledge experts periodically. An alternative was further developed to experiment with a centralized rehabilitation patient scheduling solution.

The results indicated that there were no major bottlenecks or discrepancies with the proposed solution. The composite labor cost decreased by 30% with a 60% increase in patient throughput.

Return to Top

Architecture Analysis and Modeling

Project - Architecture Analysis and Modeling for Banfield

Client - Banfield of Medical Management International, Inc. - Portland, OR

Project Statement

Banfield, a subsidiary of Medical Management International, Inc., had a requirement to conduct an analysis of its existing system architecture. System architecture is defined as the servers, clients, network, data and applications components that make up the system. The existing architecture consisted of dialup lines and local area networks that supported the store FoxPro applications and the reporting to headquarters in Portland, OR. The analysis had to identify key information to upgrade and evolve the architecture based on future plans with Pet-II and to share networking assets with PetSmart. Pet-II is the next generation of applications needed to better support current and future business plans with the Banfield information systems in the hospitals and headquarters.

CACI Services Involvement

The objective of this engagement was for CACI to provide analysis and planning support to the Banfield architecture and application team to better determine the requirements the new architecture. The analysis included support in identifying and validating possible choices and impacts of varying client/server application and data base architectures within the new Pet-II information system and any associated networking issues. Since the application and data base architectures of Pet-II may drive performance and bandwidth requirements on the networks and servers in the future, care was taken in the planning for network upgrades to meet those capacity requirements.

Deliverables

CACI provided Banfield with four COMNET III "What-If" simulation models and a written report documenting findings from the model/personal research.

Results

Performance results from the four "What-If" simulation models indicate planned infrastructure can support Pet II at a load level up to three times the expected average load. Simulation models also indicated the 56kbps V.90 modem uplink being heavily used in large hospitals (54%) during an average hour.

Return to Top

Point of Dispensing (POD) Model

Model Description

Staffing of Points Of Dispensing (PODs) of medications in a biological event requiring mass prophylaxis is problematic because of expected outage due to illness among staff members and general shortages of medical personnel. Simulation analysis can assist with balancing staffing among POD functions for maximal dispensing throughput with the available staff.

In the POD simulation model four types of patients are generated

1. Type E (Express)
2. Type S (Screening needed)
3. Type C (Counseling needed)
4. Type T (Translation needed)
5. Type D (Divert to hospital)

Each call type has its own rate of occurrence in the population, which is set at a generate activity inside the GenPatients simulation activity.

The POD is staffed by various types of medical resource persons

• Screeners, who determine what medications can be given to the patients
• Consultants, who work with difficult patients to establish medical needs
• Translators, who screen patients in languages other than English
• Dispensers, who dispense prophylactic medications
• Medical Evaluators, who divert sick patients to hospitals

The numbers of each type of staff member type available in the model are model parameters that the user can change each time the simulation is started. This is done by means of a dialog box that opens automatically. The user changes the model staff parameters from that dialog. Some typical staff counts are represented in the summary of five model runs shown in the table below

For security reasons, no more than a specified number of patients can be in the POD at any time. Others must remain queued up outside the POD. Simulation study can be used to estimate line lengths at POD activities in relation to the staff complement. The POD model includes a number of relevant real time plots that can be used to evaluate the line lengths in the dispensing queues. It can be seen from the Dispense Wait graph for instance, that the number of patients in line for an allocation of seven dispensing persons is often between five and 20 persons waiting. Increasing the number of dispensers to nine reduces this waiting line to a more acceptable value of one or two patients.

The objective is to achieve maximum patient throughput over the 48-hour operation of the POD while using the minimum staff. The model patient throughput and POD processing times for the four parameter sets above are shown in the following table

From this study it could be concluded that the parameter staffing of Model 3 or Model 4 gives the greatest throughput for the staff required and allowable line length. Note that the additional Medication Dispensing staff member between Models 3 and 4 does not contribute significantly to throughput. This staff member could be transferred to another activity or another POD.

This model was built using advanced features available in SIMPROCESS Professional.

Demonstration Model - Point of Dispensing (POD) (POD09D.spm)

Return to Top