Skip Navigation LinksHome > July/August 2001 - Volume 31 - Issue 7/8 > Simulation Software: Engineer Processes Before Reengineering
Journal of Nursing Administration:

Simulation Software: Engineer Processes Before Reengineering

Lepley, Cyndi J. PhD, RN

Free Access
Article Outline
Collapse Box

Author Information

Author affiliation: Premier, Inc., Fort Worth, Tex.

Address for correspondence and reprint requests to: Cyndi J. Lepley, PhD, RN, PO Box 16094, Fort Worth, TX 76132 (

Collapse Box


People make decisions all the time using intuition. But what happens when you are asked: “Are you sure your predictions are accurate? How much will a mistake cost? What are the risks associated with this change?” Once a new process is engineered, it is difficult to analyze what would have been different if other options had been chosen. Simulating a process can help senior clinical officers solve complex patient flow problems and avoid wasted efforts. Simulation software can give you the data you need to make decisions. The author introduces concepts, methodologies, and applications of computer aided simulation to illustrate their use in making decisions to improve workflow design.

Simulation is “any analytical method meant to imitate a real-life system, especially when other analyses are too mathematically complex or too difficult to reproduce.”1(p118) It combines spreadsheet concepts, flow diagramming methodology, and the benefits of a computer automated design system with the manager’s intuition. Put more simply, computer simulation allows you to build a model of the real system. “It is an experiment, that lets you evaluate your ideas, before you make a decision and implement a change that produces an undesirable outcome.”2(p30) Professor Jack Matson of the University of Michigan calls it “intelligent fast failure,” learning from mistakes, before they happen. 2(p30) The application of simulation is the application of scientific experimentation, which results in new knowledge and maybe even new information we were not expecting. It is useful when the real thing cannot be studied.

In a hospital system or in a clinic setting, complexity, variability, randomness, and interdependencies within the system, make processes difficult to understand and evaluate. Complexity of a system, combined with the political or personal interests of others, compounds decision making for nurse executives today. The task of changing deep rooted traditional processes to new ones that are more efficient leads to numerous “creative” suggestions from staff members, physicians, and administrators. Everyone has ideas about what will work best. Their motives for change are not always without political implications; often, they come from a desire to further personal interests. If efficiency must be improved and change is imminent, simulation of day-to-day processes can enlighten the nurse executive’s decision-making process. When you need action in the face of uncertainty, SIMULATE.

Back to Top | Article Outline

Current State of Simulation

The fields of simulation and modeling applications are expanding rapidly as software becomes easier to use. For example, the Monte Carlo probability model is used by Wall Street to study the behavior of stock prices; architects can plan for crowd reactions by designing a building using simulated traffic management strategies and pedestrian flow models to help direct the flow of people 3; athletes at all levels use simulation games to practice their skills; weather reporters use simulations to predict weather; hair stylists and plastic surgeons use simulations to demonstrate outcomes of intended hair cuts or surgeries.

Back to Top | Article Outline

Business Simulation

The use of simulation has rapidly expanded in the business world for purposes of problem solving and making decisions. One researcher even suggests that simulation software will be standard on computers shipped for home and business use in the near future. Simulation is a virtual reality vehicle for problem solving in the business world. Business simulations let you create a model and test “what if” scenarios before you attempt to change or launch a new strategy.

A business simulation tool called “Project Challenge” uses objects to help business associates contend with deadlines, budgets, and various team members, who may sometimes be antagonistic. This tool helps prepare you for surprises while you are in a safe environment. For example, your surprise might not come in the form of an antagonistic coworker, but in the form of a business war, called strategy simulation. 4(p1) In business, and presumably in healthcare, “simulation kills bad ideas fast, and breathes life into good ideas that might be politically unpalatable.”4(p4)

In healthcare, educating students is expensive. Animal laboratories used to teach physiology and hands-on skills such as incubation techniques are costly to maintain. As a result, many schools use simulation based training to help students become active learners rather than become passive recipients of information. For example, surgical training simulations help surgeons “feel” the force of a knife as they cut through skin to avoid injuring a major organ or cutting a blood vessel. Simulations also help students practice making decisions to learn how their thought processes and actions affect outcomes. Just as simulation for a business administrator provides models that allow control over a system, training simulation tools provide advantages for healthcare or nursing students by focusing their learning on the problem, allowing them control of the situation and the speed of their educational experience. Most important, these tools prevent the patient from being exposed to poor decision making.

Back to Top | Article Outline

Why Use It Now?

People are becoming more familiar with statistics packages and control charts, and are building flow models of processes. Thus, it seems realistic to expect software simulation to be used by mainstream society. The results of a 1998 study show that the use of animated flow charts, statistical packages, and computer simulation may soon become mainstream technology, with simulation software sold in all new computers. 5 Simulation products make it practical to redesign processes, even on a day-to-day basis if necessary.

Computer flow chart modeling techniques for simulation (for example, ProcessModel™ [Process Model, Inc., Provo, UT]) reproduce the behavior of a process graphically and logically, so constituents can see the pictures and “find-by-looking” at the bottlenecks that affect resource use. When we can see what might happen in an activity, we are more likely to believe that what is being projected as an outcome, could really happen! Because computers and software can imitate behaviors in real systems using the models we construct, it is possible to imitate an epidemic in a community or a new process on a patient care unit, which otherwise cannot be studied easily. A simulation study manipulates a model and tests ideas and newly designed processes before you attempt a change or launch a new strategy. You learn something about a proposed change from experiments that represent the new system before you change its design.

Management science combines the use of spreadsheets, computers, statistics, and mathematics to solve problems. The system within which a problem resides may be so complex that it is impossible to understand the effect of an intervention because of the existence of the uncontrollable parts and the interactions of the controllable parts. 6 When the problem is complex, decisions are easier to make if data and a picture are available to describe what the changed model will look like or how it will perform. A computer simulation, as it runs, gathers data about an object in the system; these data are based on the real world example. The aim is to understand the object’s performance within the system to predict, change, and control the behavior of a system. 2(p30) Once you understand what is going on, the future can be designed to meet your objectives. The road map does not have to lead to chaos. Simulation is about risk analysis and preventing chaos before you implement a disaster. You can change a risk to a predictable outcome that is the result of your own creation for “what if” situations. You predict the outcomes to your own creations for “what if” situations and you make the best decision based on sound data.

When you want to control an unpredictable situation, simulation with a spreadsheet model or with an animated flow chart model can predict outcomes and identify risks of various scenarios in a system. The models find the solutions of the problems. Modeling is the process of establishing interrelationships between important entities of a system. Before modeling, it is necessary to express the expected outcomes in terms of goals and performance criteria, so that the model can help you find the constraints. Using animated simulation applications to create flow chart pictures of processes helps you find the interdependency within the system constraints. 7(p40)

Back to Top | Article Outline

Simulation in Hospitals

The administrative decisions that one makes in healthcare are well suited for computer simulation. Both spreadsheets and animated flow charts can help identify flow problems, which could be caused by a lack of resources or outdated policies. Studies might result in changes such as alternative admissions policies, revised scheduling of multiple operating rooms, management and control of outpatients in a labor and delivery unit, control of patient gridlock in the emergency department, or an inpatient unit.

In healthcare systems, patients rightfully expect us to focus on their specific needs during their hospitalization. With simulation, hospital administrators can test ways to design patient care delivery systems to achieve high efficiency at low costs, and still meet patient expectations. Examples of problems are built and simulated as static models (time independent) or as dynamic situations (time independent). 8(p128) The static model shows the relationship between entities and attributes if the system is in equilibrium. In such a model, the observer can see how a change in the value of an attribute affects the system. However, with this model, the observer cannot begin to understand how or why the change occurred. As time passes in a system, a dynamic model represents the sequence in which events occur and describes the flow of an entity through the system. 9 “The goal of a system simulation model is to reproduce the activities that control the flow of entities and the logic by which events occur over time.”9(p146)

Back to Top | Article Outline

Managing Patient Flow Study: Length of Stay and Bed Blocking

Hospital problems are best understood if presented as dynamic models, models that change as a function of time. A common problem in hospitals that must be resolved is managing the flow of patients between units, from the ICU (intensive care unit) to a medical surgical unit, and from the emergency department to an empty bed. This problem can be resolved with simulation using the expected length of stay, the average daily census, and the average number of scheduled admissions.

One hospital used flow modeling to study how patient flow, physician referrals, and patient volume affected bed availability on an inpatient unit and flow to a long term care setting. 10 The research project reviewed bed occupancy data and the average daily census in an acute care setting. It developed a model that forecast the bed requirements based on the average length of stay, estimated daily admissions, beds in each units, expected daily discharges to long term care and the average number of patients waiting to be placed in long term care.

After the model was created, the researchers considered the problem a queuing system issue and used “what if” analyses to help them understand the bed requirements. With an increased understanding of their system, they developed a new method for inpatient management in the general hospital.

Back to Top | Article Outline

Patient Flow in a Simulated Clinic Model

Simulation can show how the users of processes compete for resources. In the model shown (Figure 1), patients enter a clinic system to move through various stations on their way to see a physician. Resources consist of the admission clerk, laboratory and X-ray technicians, nurses, and physicians. The system must produce enough patients to keep the employee resources productive, or the clinic will be viewed as inefficient and costly. To forecast the best scheduling pattern for patients, a simulation model (using ProcessModel™) is under design by a faculty member at the University of Alabama Birmingham to find the best times to schedule patients. The best arrival time allows patients to arrive at the examination room with their completed test results when the doctor is ready to see them. The simulation will identify where delays or constraints exist in the process of moving patients from admissions, to lab or to the X-ray department, and then to an exam room. If the system behavior does not match productivity expectations, the simulation model will be manipulated to reach the desired results, thus presenting opportunities to eliminate or decrease non-value-added activities.

Figure 1
Figure 1
Image Tools
Back to Top | Article Outline

Medication Errors

Wakefield 11 reported a spreadsheet simulation study to estimate the number of actual and perceived rates of MAEs (medication administration errors) in a hospital setting. Administering medications is a high volume activity that is an important aspect of nursing care. Medication doses are administered “at several times during the day, by different levels of nurses, with changing staffing levels and skills, who care for a fluctuating number of patients with highly variable illness severity.”11(p73) The errors can be related to legibility of the original order, transcription, prescribing errors, dispensing errors, and administration errors. The simulation, presented as results of mathematical computations, demonstrated there was a potential for many unreported medication errors. This study also showed a perception that many medication errors are unreported. The approach, although focused on understanding nurses’ perceptions of actual versus reported medication errors, indirectly emphasized the value of studying a system of activities that affect patient care and patient outcomes. Relying on reported data may be inadequate; therefore, further study may provide additional insights and innovative solutions.

Back to Top | Article Outline

Medication Ordering and Delivery

Actual administration of a drug is the last step in a series of activities that take place after the physician orders a medication. Sending the order to the pharmacy from the patient care unit and delivering the medication from the pharmacy to the patient care unit are the responsibilities of several people. Completing each step correctly and in a timely manner, contributes to prevention of errors. In one project, described in the following paragraphs, the coordination of the processes centered on the unit secretary’s ability to transcribe orders in a timely manner.

Back to Top | Article Outline

Building a Simulation

Step 1: Problem Identification

Physicians and staff members of a midwestern hospital identified problems in the medication delivery process. 12 Physicians told administrators that it took longer than 4 hours to get orders transcribed and medications administered to their patients. Their constant complaints described problems in the system. The physicians and senior administrators agreed to study the process and perform the following actions:

* Identify where and why delays occurred in the system.

* Decrease the time required to transcribe orders.

* Decrease the time required to deliver medication to the unit.

Nurses, secretaries, pharmacists, and technicians reviewed the process and identified each action step. The committee was charged with finding the cause for the amount of time it took to transcribe orders and deliver medications to the patient care units.

They had to find key steps that were not performed systematically, or those that were performed inconsistently. They found the following:

* Two to four hours passed before physician orders were transcribed.

* No specific person was accountable for transcription.

* Policies and procedures created bottlenecks in the process.

* The system lacked a standardized approach among patient care units for informing the pharmacy about medication orders.

Back to Top | Article Outline
Step 2: Conceptualize the Process

The results of any improvement project must help users to see and understand its relevance to their own daily work processes. Otherwise, cooperation is poor or lacking entirely when it is time to make changes suggested by the committee. To stimulate reflection about problems, a committee might ask these questions:

* What triggers the start of this process?

* Why do we need to change the process?

* What prevents the patient from receiving medication in a timely manner?

* What products do you provide to the patient as part of this process?

* Who are the stakeholders and what are their requirements?

Back to Top | Article Outline
Step 3: Create and Verify the Model

In this example, discussions led the improvement team to create a crossfunctional flow chart (Figure 2) that modeled their system. The chart showed that once orders were written, a minimum of 8 hours to 12 hours could elapse before the medication would be delivered back to the patient care unit. The stakeholders began work on a new process to meet their objectives. First they verified that the medication ordering model was valid and added actual times to the flow process, as shown in the “clock column” of Figure 2. This particular project did not use animated simulation to define the problems. Thus, one could only imagine the flow of the medication order through the system, not view an animation, from the time the order was written, to the transcription, to the time it was picked up on the unit, delivered to pharmacy, and packaged for delivery to the patient care unit. Using this type of flow chart, you can only guess where the bottlenecks might be, and the design of this process improvement stopped here. If simulation had been incorporated as part of continuous improvement, the committee could have proceeded to steps four and five. Instead, the organization made changes in policies, use of resources, and technology to increase physician satisfaction with medication delivery to their patients.

Figure 2
Figure 2
Image Tools
Back to Top | Article Outline
Step 4: Design Experiments and Test Results

When simulation is incorporated as part of a continuous improvement a ProcessModel flow diagram (Figure 3) is developed.

Figure 3
Figure 3
Image Tools
Back to Top | Article Outline

Model Explanation

This model’s parts consist of the following:

* An entity: the physician’s order;

* Entity attributes: Is it stat? Yes or no.

* Activities: All activities occurring after the order is written.

Two processes exist for each type of medication order. The entity (physician order) flows through the system, depending on whether it was or was not urgent. From the automated view of the process, if the nurse signs the stat order 5 minutes after the order is written, statistics suggest that the medication might be delivered in less time than if the registered nurse always used this approach for signing the orders.

As part of step four, when you design experiments and test the results, you create and answer the “what if” questions. What happens if you change a policy, change staff responsibilities, add or delete resources, or add new technology? Any of these changes can lead to interesting simulation experiments with results that find bottlenecks and help make decisions, without affecting patients during the experimental stages. The following are some “what if” scenarios:

* Would an additional clerk’s salary create time savings that would justify the added salary?

* Would automated software for physicians decrease the time required for completing order transcription and decrease errors?

* Does the time that the physician orders spend in a queue suggest that a policy change is necessary?

* If the pharmacist were on the unit to receive the order, prepare it, and possibly administer it to the patient, would the time from “order written to medication administered” be decreased?

Back to Top | Article Outline

Data Needed to Conduct Such a Study

These experiments can be simulated in any hospital, with adaptations specific to that system. Someone designs the flow chart, collects, and records the data that will eventually be put into the simulation model. It is helpful as the project begins to create a project notebook in which one keeps all data and change actions taken throughout the analysis of the current situation and the design of the new process. The data one collects for the pharmacy project would include information needed to create and simulate the “as is” model. The result of this step tells one what is happening currently. The redesigned experiment will change the outcome data to show how the future could look if the process were changed. Examples of data needed for this model are given below.

* Entities: Items or people being processed (ie, products, documents, customers). In this example, the physician orders are the entities. One must determine how many orders are written by time of day.

* Entity attributes: Values associated with each entity that may indicate the entity’s size or condition. In this example, the order’s status, regular or STAT, is its attribute. One must determine how many orders are regular or STAT.

* Activities: The tasks performed on entities, such as assembly or document approval. An activity is defined in terms of the activity time and the resource requirements. In the example, the order must be signed and transcribed by a clerk and an RN; this uses resource time. Then the order is moved through the hospital to the pharmacy by a technician who gives it to the pharmacist for medication preparation. To run the simulation, one needs to know the amount of time used by each resource to complete their activity.

* Storages: The storages are the waiting areas where the entity can wait or is held until it is processed. In the example, the orders wait to be processed until two people sign them. In this example, the charts are placed at a certain location on the desk, until signed and transcribed. Some of the data collected would include the amount of time the order waits from when it is written, to when it can be transported by the pharmacy technician.

Back to Top | Article Outline
Step 5: Draw Conclusions and Implement Changes

With a simulation tool, the statistics from the simulation predict the outcomes of each experiment. The experiments can incorporate multiple changes, creating new possibilities with different results each time. Outcome statistics, because they will differ for each organization, suggest the “best practice” in that organization. If one finds that outcome statistics are different on each patient care unit in the same hospital, processes may need to be standardized. If the approach on the patient care units can be standardized, other departments could standardize their processes, thus improving efficiency and most likely decreasing costs. A well executed simulation project will provide completed models of a current process and after analysis of the data, suggest alternatives that will decrease resource use, improve timeliness, and decrease errors. A successful simulation project should not only result in improvements in the process studied (ie, reduction in costs and cycle time), but also point out other areas of potential improvement within the same system.

Back to Top | Article Outline

Final Thoughts

Both spreadsheets and flowcharts can produce simulated data for decision making. If internal resources to develop an accurate model (both time and staff) are lacking, it may not be possible to improve a process with simulation. For example, software prices range from a few hundred dollars to $3,000–$20,000. Someone must be trained to use and understand the logic of the software, have the time to gather the data, and create and validate the model. The development of a simulated process can take days, even weeks. Although it can be expensive and time consuming to adopt simulation as a major tool for analyzing and changing processes, it can be more expensive to redesign after implementing the wrong process change. Figure 4 provides a partial listing of websites that provide useful information about simulation and simulation products.

Figure 4
Figure 4
Image Tools

In a recent study, CEOs (chief executive officers) were asked to describe the trends likely to affect business in the future. More than 51% reported that one trend would be an ongoing need to measure and analyze organizational processes. If the cost and cycle time of a process is reduced by effectively managing product and service design and delivery, an organization will remain competitive. 13(p14) The CEOs also reported that a large gap exists between management’s current competency and desired competency in several major trends, one being competency in the ability to reduce cost and cycle time of processes. 13(p14)

To decrease this gap, managers must understand how to use a simulation model to create experiments that vary the outputs of processes. Various types of output data will direct them through an analysis of the steps of a complex process and toward fact based decision making. Results of simulation produce statistical data to help managers make decisions about change using facts, not intuition. Using statistical data they can examine processes, even optimize them, to achieve “higher levels of performance, eliminate errors, and reduce cycle time and costs.”13(p26) “Gaining the luxury of playing with hypothetical “what if” scenarios on a computer model without disturbing an actual system can make a significant impact on strategic decisions.”14(p47)

Managers often claim that a particular situation is so complex and unpredictable that it cannot possibly be modeled. Managers must become believers and obtain useful results quickly from a model they help to build. The resulting improvements in efficiency and effectiveness that occur from their improved processes enable managers to contribute to an organization’s ability to perform better, faster, and cheaper.

Another key aspect of survival in the future is agility or the capacity for rapid change and flexibility. Simulation shows the way to create simpler processes for an infrastructure. It even provides clues to help understand the risks of innovative ideas. Many companies across the country and outside of the United States have adopted simulation to help them change quickly from one process to another. Can healthcare afford to be far behind?

Simulation’s logic and statistical output will not always be the fastest way to solve a problem, but they will produce proof of a concept or an innovation that can be taken to key decision makers and stakeholders. Simulation tools that model flow charts help decision makers visualize the location of bottlenecks, making it easier for the change agent to explain the problem, show its causes, and sell the solutions. If one makes decisions using intuition alone, the assumptions and results will differ from person to person. This can be disastrous. Once disaster occurs, it is difficult to explain the rationale behind the original recommendations and decisions. One can minimize the costs of mistakes from intuitive decisions and conduct real world experiments that perform mathematical analyses with simulation tools. The chances of successful change increase with a higher level of performance when cycle time and costs are reduced.

Back to Top | Article Outline


1. Walker R. Future Farmers of America. Fast Company. 1997;7 (February):118.

2. Barrett D. The Paradox Process: Creative Business Solutions Where You Least Expect to Find Them. New York, NY: American Management Association; 1998.

3. Low D. Statistical physics: Following the crowd. Nature. 2000; 407 (9–28); 465–466.

4. Carbonera P. Game Over. Available at:http://www.fastcompany. com/online/06/gameover.html. Accessed 9/1/00.

5. Garnett J. The Last Word on Simulation. IIE Solutions. Available at: Accessed 1/99.

6. Zeigler BP. Multifaceted Modeling and Discrete Event Simulation. New York, NY: Academic Press; 1984.

7. Wyland B, Buxton K, Fuqua B. Simulating the supply chain. IIE Solutions. Available at: Accessed January 2000: vol.32, no.1.

8. Graybeal WT, Pooch UW. Simulation: Principles and Methods. Boston: Little, Brown and Company; 1980.

9. Evans JR, Olson DL. Introduction to Simulation and Risk Analysis. Denver, CO: Decisioneering, Inc.

10. El-Darzi E, Vasilakis T, Millard PH. A simulation modeling approach to evaluating length of stay occupancy, emptiness and bed blocking in a hospital geriatric department. Health Care Management Science. 1998; 1: 143–149.

11. Wakefield DS, Wakefield BJ, Borders T, et al . Understanding and comparing differences in reported medication administration error rates. Am J Med Qual. 1999; 14 (2); 73–79.

12. Lepley C. Prescription for success. Simulation Sucess. Available at: Accessed 4/2000.

13. Blazey ML, Ettinger JH, Grizzell PL, Janczak LM. Insights to Performance Excellence in Health Care 2000: An Inside Look at the 2000 Baldrige Award Criteria for Health Care. Milwaukee, WI; ASQ Quality Press: 2000.

14. Profozich D. Managing Change With Business Process Simulation. Upper Saddle River, NJ; Prentice Hall PTR; 1998.

© 2001 Lippincott Williams & Wilkins, Inc.