Search

Spine surgeries tend to be lengthy (mean time of 4 hours) and highly variable (with some surgeries lasting 18 hours or more). This variability along with patient preferences driving scheduling decisions resulted in both low operating room (OR) utilization and significant overtime for surgical teams at Mayo Clinic. In this paper we discuss the development of an improved scheduling approach for spine surgeries over a rolling planning horizon. First, data mining and statistical analysis was performed using a large data set to identify categories of surgeries that could be grouped together based on surgical time distributions and could be categorized at the time of case scheduling. These surgical categories are then used in a hierarchical optimization approach with the objective of maximizing a weighted combination of OR utilization and net profit. The optimization model is explored to consider trade-offs and relationships among utilization levels, financial performance, overtime allowance, and case mix. The new scheduling approach was implemented via a custom Web-based application that allowed the surgeons and schedulers to interactively identify best surgical days with patients. A pilot implementation resulted in a utilization increase of 19% and a reduction in overtime by 10%.

Purpose – Parallel processing of regional anesthesia may improve operating room (OR) efficiency in patients undergoes upper extremity surgical procedures. The purpose of this paper is to evaluate whether performing regional anesthesia outside the OR in parallel increases total cases per day, improve efficiency and productivity. Design/methodology/approach – Data from all adult patients who underwent regional anesthesia as their primary anesthetic for upper extremity surgery over a one-year period were used to develop a simulation model. The model evaluated pure operating modes of regional anesthesia performed within and outside the OR in a parallel manner. The scenarios were used to evaluate how many surgeries could be completed in a standard work day (555 minutes) and assuming a standard three cases per day, what was the predicted end-of-day time overtime. Findings – Modeling results show that parallel processing of regional anesthesia increases the average cases per day for all surgeons included in the study. The average increase was 0.42 surgeries per day. Where it was assumed that three cases per day would be performed by all surgeons, the days going to overtime was reduced by 43 percent with parallel block. The overtime with parallel anesthesia was also projected to be 40 minutes less per day per surgeon. Research limitations/implications – Key limitations include the assumption that all cases used regional anesthesia in the comparisons. Many days may have both regional and general anesthesia. Also, as a case study, single-center research may limit generalizability. Practical implications – Perioperative care providers should consider parallel administration of regional anesthesia where there is a desire to increase daily upper extremity surgical case capacity. Where there are sufficient resources to do parallel anesthesia processing, efficiency and productivity can be significantly improved. Originality/value – Simulation modeling can be an effective tool to show practice change effects at a system-wide level.

Patient appointment booking, sequencing, and scheduling decisions are challenging for outpatient procedure centers due to uncertainty in procedure times and patient attendance. We extend a previously developed appointment scheduling model to formulate a model based on a two-stage stochastic mixed integer program for optimizing booking and appointment times in the presence of uncertainty. The objective is to maximize expected profit. Analytical insights are reported for special cases and experimental results show that they provide useful rules of thumb for more general problems. Three solution methods are described which take advantage of the underlying structure of the stochastic program, and a series of experiments are performed to determine the best method. A case study based on an endoscopy suite at a large medical center is used to draw a number of useful managerial insights for procedure center managers.

In this article, we report on the implementation of a computerized scheduling tool to optimize staffing for patient transport at the Mayo Clinic. The tool was developed and implemented in Microsoft Excel and Visual Basic for Applications and includes an easy-to-use interface. The tool allows transport management to consider the trade-offs between patient waiting time and staffing levels. While improved staffing efficiency was a desire of the project, it was important that patient service quality was also maintained. The results show that staffing could be reduced while maintaining historical patient service levels.

In August 2010, the Third Annual Mayo Clinic Conference on Systems Engineering and Operations Research in Health Care was held. The continuing mission of the conference is to gather a multidisciplinary group of systems engineers, clinicians, administrators, and academic professors to discuss the translation of systems engineering methods to more effective health care delivery. Education, research, and practice were enhanced via a mix of formal presentations, tutorials, and informal gatherings of participants with diverse backgrounds. Although the conference promotes a diversity of perspectives and methods, participants are united in their desire to find ways in which systems engineering can transform health care, especially in the context of health care reform and other significant changes affecting the delivery of health care.

We report on the use of discrete event simulation modeling to support process improvements at an orthopedic outpatient clinic. The clinic was effective in treating patients, but waiting time and congestion in the clinic created patient dissatisfaction and staff morale issues. The modeling helped to identify improvement alternatives including optimized staffing levels, better patient scheduling, and an emphasis on staff arriving promptly. Quantitative results from the modeling provided motivation to implement the improvements. Statistical analysis of data taken before and after the implementation indicate that waiting time measures were significantly improved and overall patient time in the clinic was reduced.

PurposeThe purpose of this paper is to report on the development of a qualitative systems model developed to understand why average emergency department (ED) length of stay (LOS) was rapidly increasing while the number of ED visits was relatively constant. The paper's focus was to identify systemic causes for poor patient flow so that the model could then be used to evaluate improvement options using a more complete view of the causal structure for the ED delays.Design/methodology/approachIn this case study, a disciplined system dynamics approach was used that included development of a dynamic hypothesis, causal loop and stock and flow diagramming, interviews with system experts, and data collection and analysis.FindingsResults support the dynamic hypothesis that an aging population and shortages of resources to treat chronically ill patients (among other dynamics) were causing longer average LOS. Older and sicker patients were consuming more ED resources and causing less acute patients to leave without being seen or to avoid visiting the ED in the first place. In essence, the ED was acting as a safety valve for the wider health care system as many parts of this wider system became overloaded.Practical implicationsOwing to the systemic causes for the patient treatment delay problem in the ED, simple local solutions are unlikely to be effective. The system model can be used as a basis to understand the underlying dynamics of the systemic causes for poor patient flow and identify robust and long‐term solutions.Originality/valueThe paper presents a process for developing a dynamic model to engage the various participants in a health care system in understanding the causes for delays and poor patient flow. The modeling approach can be used as a means for health care managers/administrators to identify improvement options that address the systemic problems.

The early/tardy problem is one of the most vexing pieces of the complex production scheduling decision process. So far most of the research has been on single-machine environments. Hence, we considered the weighted early/tardy scheduling problem in a simulated dynamic multimachine job shop. We analyzed controlled job-release and dispatch rules using time and cost information at a variety of stationary and nonstationary utilization rates, due-date allowances, and early/tardy cost levels. We found a newly developed method for controlling the release for all job operations using early/tardy cost information, superior to other release mechanisms overall in both our stationary and nonstationary analyses. We found immediate release useful at many high utilization conditions and a gateway-only release method best in many low utilization conditions. A modified version of a single-machine early/tardy dispatch method was clearly superior to the dispatch rules for almost all the simulated shop conditions.

Many disasters have occurred because organizations have ignored the warning signs of precursor incidents or have failed to learn from the lessons of the past. Normal accident theory suggests that disasters are the unwanted, but inevitable output of complex socio-technical systems, while high-reliability theory sees disasters as preventable by certain characteristics or response systems of the organization. We develop an organizational response system called incident learning in which normal precursor incidents are used in a learning process to combat complacency and avoid disasters. We build a model of a safety and incident learning system and explore its dynamics. We use the model to motivate managers to implement incident learning systems as a way of moving safety performance from normal accidents to high reliability. The simulation model behavior provides useful insights for managers concerned with the design and operation of incident learning systems. Copyright © 2006 John Wiley & Sons, Ltd.

Managing production and inventory for large-scale continuous processing plants is a key to success in the chemical process industry, particularly in the production of commodity polymers such as polyethylene and polypropylene. When setting up a production schedule, planners must consider the effects of off-grade production and production rate penalties, which are typically sequence-dependent as products are transitioned from one to another, as well as inventory holding costs and capacity constraints. We call this the continuous economic lot sizing and scheduling problem (CELSP) to distinguish it from closely related problems found in discrete part production industries. We present a formulation that addresses the particular aspects of the CELSP and apply the proposed mathematical modeling approach to several plants of a chemical processing company. This company was concerned with ensuring the plants were carrying proper amounts of inventory and with evaluating the number of products produced at each plant. Our results show actual overall inventory levels were close to the levels suggested by our model and therefore company plans for further inventory reductions would not be appropriate. However, the company should carefully consider the addition of new products to current plant product slates. Due to the effects of product transitions and inventory levels, even adding products with significant contribution margins may negatively affect the financial performance of the plant.

Time waiting for service is a major concern for consumers, and excessive waiting for a pre‐scheduled appointment is especially annoying. This is an on‐going problem because appointment scheduling is a challenging task, mainly due to the uncertainties associated with service times. Prior studies have focused mainly on a single scheduling period (i.e. either a morning or afternoon); this paper uses a more realistic model that represents an on‐going, multi‐period scheduling environment where clients can be scheduled days or even weeks into the future. Two main objectives will be considered; the best scheduling rule to use in a multi‐period environment, and the best placement of appointment slots that are left open for urgent clients. Both of these have been studied in a single period environment, and results here will be compared to those. It will be shown that in some cases earlier findings from the one‐period environment are robust and perform well in a multi‐period environment, while in other cases the one‐period findings do not apply.

We extend past research on the economic lot scheduling problem to address some of the limitations of the earlier work. In particular we develop mixed integer programming formulations with the assumption of a production precedence sequence. When evaluated over a variety of problems from the literature, it is clear that finding a schedule that minimizes costs is not trivial. However, with the use of an effective sequencing heuristic and appropriate formulations, optimal schedules can often be found.

Service managers are continually challenged with balancing customer demand and service capacity. Recent studies have raised awareness of various demand and capacity management practices available to services, but little numerical work has been done to identify how these decisions work together and how they relate to one another. For instance, reducing prices may attract customers during a slow period, but the extent of impact this should have on cross‐training staff is not clear. A simulation based on theoretical and empirical insights explores the impact of various decisions on profitability and operations. The decisions modelled include the impact of: automation, customer participation, cross training employees, informing customers about the operation, and others. It is shown that demand and capacity decisions do indeed impact on each other – sometimes in ways that are not initially obvious. Results provide useful thought‐starters for service managers striving to improve their operations.

We develop a method for finding the economic sampling plan that minimizes expected total costs, subject to a constraint on labour time. We explicitly recognize the labour time required for sampling, investigation, extra processing effort per unit due to operating the process out of control, and the time to put the process back into control. These times, as well as the total production time, are frequently constrained by total available labour. With this constraint, we use a Lagrangian relaxation method to find the best sampling plan. Results of a numerical experiment show that the labour-constrained sampling plans and costs can differ substantially from the unconstrained solution. However, much of the cost penalty when labour is tightly constrained can be reduced with relatively small increases in total available labour time.

AimWhile early warning scores (EWS) have the potential to identify physiological deterioration in an acute care setting, the implementation of EWS in clinical practice has yet to be fully realized. The primary aim of this study is to identify optimal patient-centered rapid response team (RRT) activation rules using electronic medical records (EMR)-derived Markovian models.MethodsThe setting for the observational cohort study included 38,356 adult general floor patients hospitalized in 2011. The national early warning score (NEWS) was used to measure the patient health condition. Chi-square and Kruskal Wallis tests were used to identify statistically significant subpopulations as a function of the admission type (medical or surgical), frailty as measured by the Braden skin score, and history of prior clinical deterioration (RRT, cardiopulmonary arrest, or unscheduled ICU transfer).ResultsStatistical tests identified 12 statistically significant subpopulations which differed clinically, as measured by length of stay and time to re-admission (P

Although profitability is generally increased when services do well at matching demand and capacity, service managers continue to struggle in this area. This article summarises major demand and capacity management literature from both operations and marketing fields with the view to combine the knowledge into a pragmatically based framework. The primary focus is on services that cannot schedule customers (e.g., banks, supermarkets), since these provide the greatest challenge for demand management. Narrowing this further, the article gives preference to work that has information useful to promote future modelling efforts.

Consider an organization whose capability to produce an item and whose customer demand are both stochastic. In such a context “take-or-pay” contracts can be attractive. Under such a contract the organization agrees to purchase from a supplier a fixed quantity per period over a specified number of periods. Simulation is too slow an analysis approach for the typical dynamic negotiation situation. We use a Markovian approach to create a tool that negotiators can use to evaluate the expected cost of a proposed contract, considering the stochastic demand and all relevant cost components. The approach is fast enough to use in real time, and yields accurate (sometimes exact) results.

Clients of services expect short waiting times and servers desire short periods of non-productive time. One of the areas where this is most important is appointment scheduling systems. Recent research has indicated that using information about clients’ service time variability can simultaneously reduce waiting times and server idle time. In this study, a more realistic, dynamic appointment-scheduling environment is developed and used to analyze several scheduling rules. Additional complexities considered in this study include: continuously distributed service-time variances, special client appointment requests, and appointment-scheduler uncertainty. Results show that rules using client-variance information are still best at minimizing waiting time and idle time with the additional complexities. In fact, these rules perform best when client variance is large. However, on measures related to clients requesting specific appointment slots the results are not as clear cut. A key factor for these measures is the distribution of the desired slots. When the desired slots are near the end of the appointment scheduling period, traditional rules like first-call-first-appointment perform better on client appointment request measures.

In today's fiercely competitive marketplace, firms are looking for ways to improve their profitability. Computer simulation studies comparing group technology/cellular manufacturing (GT/CM) to job-shop layouts addressed this topic. However, each study ignored some of the most basic operating conditions said to give GT/ CM its advantages: this simulation study incorporates some of those operating conditions like human resource issues, such as learning and labour constraints, in a comprehensive setting. The results indicate that GT/CM significantly outperforms the job shop layout in almost every environmental setting.

A common approach used in quality or process improvement is cause-and-effect analysis. It involves identifying a (large) number of possible causes of an undesirable effect or problem, and then choosing the sequence of investigation of the possible cause..

Considering the strategic importance of business process improvement, it is imperative that educators, students, and practitioners be familiar with this topic. Thus, the intention of this tutorial is to provide a guiding framework for carrying out improvements of business processes. We present numerous illustrative examples, taken from our personal experiences and those of our students, as well as from the literature. An extensive reference list is provided, thus pointing the interested reader to sources of further detail on specific topics.

We used the systems engineering technique of discrete event simulation modeling to assist in increasing patient access to positron emission tomographic examinations in the Department of Nuclear Medicine at Mayo Clinic, Rochester. The model was used to determine the best universal slot length to address the specific access challenges of a destination medical center such as Mayo Clinic. On the basis of the modeling, a new schedule was implemented in April 2012 and our before and after data analysis shows an increase of 2.4 scans per day. This was achieved without requiring additional resources or negatively affecting patient waiting, staff satisfaction (as evaluated by day length), or examination quality.

Purpose Transfer of intensive care unit (ICU) patients to the operating room (OR) is a resource-intensive, time-consuming process that often results in patient throughput inefficiencies, deficiencies in information transfer, and suboptimal nurse to patient ratios. This study evaluates the implementation of a coordinated patient transport system (CPTS) designed to address these issues. Methods Using data from 1,557 patient transfers covering the 2006–2010 period, interrupted time series and before and after designs were used to analyze the effect of implementing a CPTS at Mayo Clinic, Rochester. Results Using a segmented regression for the interrupted time series, on-time OR start time deviations were found to be significantly lower after the implementation of CPTS (p < .0001). The implementation resulted in a fourfold improvement in on-time OR starts (p < .01) while significantly reducing idle OR time (p < .01). Conclusion A coordinated patient transfer process for moving patient from ICUs to ORs can significantly improve OR efficiency, reduce nonvalue added time, and ensure quality of care by preserving appropriate care provider to patient ratios.

Rising costs and increasing patient expectations have heightened the need for better coordination of health services. In this chapter we discuss the role that operations research methods such as system dynamics and discrete-event simulation can play in pursuit of such coordination. Most healthcare services are operated independently from others, although patient treatment often requires visiting several services. Also, most operations management methods have focused on improvement or design of individual services. We will highlight some of the research that considers coordination among multiple services, and present two case studies that address coordination at the healthcare system and hospital levels. We also discuss some new trends in healthcare services delivery that have significant implications for operations researchers.

The primary issue addressed in this research is how to schedule clients as they call for appointments, without knowing which “types” of clients will call at a later time. The main goal is to compare various scheduling rules in order to minimize the waiting time of the clients as well as the idle time of the service provider. Interviews with receptionists verified that they have knowledge regarding differences between clients' service time characteristics. This information is used both to differentiate between clients and to develop various scheduling rules for those clients. A simulation model of a dynamic medical outpatient environment is developed based on insight gained from the interviews and from prior research. Two decision variables are analyzed (“scheduling rule” and “position of appointment slots left unscheduled for potential urgent calls”) while two environmental factors are varied (“expected mean of the clients' service time”, and “expected percentage of clients with low service time standard deviation compared to those with high service time standard deviation”). This resulted in 30 combinations of decision variables, each tested within 15 combinations of environmental factors. By using multiple performance measures, it is possible to improve considerably on some of the “best” rules found in the current literature. The “best” decisions depend on the goals of the particular clinic as well as the environment it encounters. However, good or best results can be obtained in all cases if clients with large service time standard deviations are scheduled toward the end of the appointment session. The best positioning of slots left open for urgent clients is less clear cut, but options are identified for each of a number of possible clinic goals.

In an endeavor to broaden the application of scheduling models to decisions involving the use of a manager's time we use simulation to investigate the performance of a number of simple algorithms (including eight priority rules and a construction heuristic) in a dynamic setting with tasks arriving (randomly) and scheduling decisions being made, over time. We compare these simple methods relative to a bound that uses an adjacent pairwise interchange algorithm. We model uncertainty in task durations, and costs being incurred for early and tardy task completion ( representative of JIT settings). In addition to evaluating the efficacy of the scheduling rules and various preemption strategies (using ANOVA), we highlight the managerial implications of the effects of eight environmental parameters.

A number of articles have recently appeared proposing the use of NPV and similar economic scheduling performance measures. Much of this research has concentrated on developing methods to maximize NPV. However, it is assumed in almost all these studies that the cash flow information and cost rate parameters are all known with certainty. For confident use of NPV it would be beneficial to know the robustness of this measure to economic misspecification. This research study addresses the issue of economic information error by developing a simulation model of a jobshop that includes varying levels of information misspecification. The study considers a variety of time-based and economically oriented performance measures along with a number of classical and economic order-release and dispatch scheduling rules. Varying rates of shop utilization and due date tightness were also examined. The results show that NPV is relatively robust to information errors. Average errors must be of the magnitude of at least 40% before time-based measures such as mean flowtime and mean tardiness are financially preferable. At average error levels greater than 40%, the economic-based measures may be misguided by the poor economic information, making their schedules of less worth. However, with low utilization rates, many time-based measures are as good, or almost as good, as NPV in maximizing NPV with information error, even with low information error levels. Under such conditions, the use of the time-based measures may be preferable due to their simplicity.

Recovery beds for cardiovascular surgical patients in the intensive care unit (ICU) and progressive care unit (PCU) are costly hospital resources that require effective management. This case study reports on the development and use of a discrete-event simulation model used to predict minimum bed needs to achieve the high patient service level demanded at Mayo Clinic. In addition to bed predictions that incorporate surgery growth and new recovery protocols, the model was used to explore the effects of smoothing surgery schedules and transferring long-stay patients from the ICU. The model projected bed needs that were 30 % lower than the traditional bed-planning approach and the options explored by the practice could substantially reduce the number of beds required.

The new Accreditation Council for Graduate Medical Education (ACGME) duty-hour standards for residents and fellows went into effect in 2011. These regulations were designed to reduce fatigue-related medical errors and improve patient safety. The new shift restrictions, however, have led to more frequent transitions in patient care (handoffs), resulting in greater opportunity for communication breakdowns between caregivers, which correlate with medical errors and adverse events. Recent research has focused on improving the quality of these transitions through standardization of the handoff protocols; however, no attention has been given to reducing the number of transitions in patient care. This research leverages integer programming methods to design a work shift schedule for trainees that minimizes patient handoffs while complying with all ACGME duty-hour standards, providing required coverage, and maintaining physician quality of life. In a case study of redesigning the trainees’ schedule for a Mayo Clinic Medical Intensive Care Unit (MICU), we show that the number of patient handoffs can be reduced by 23 % and still meet all required and most desired scheduling constraints. Furthermore, a 48 % reduction in handoffs could be achieved if only the minimum required rules are satisfied.

Discrete-event simulation (DES) is an effective tool to for analyzing and improving healthcare processes. In this chapter we discuss the use of simulation to improve patient flow at an outpatient procedure center (OPC) at Mayo Clinic. The OPC addressed is the Pain Clinic, which was faced with high patient volumes in a new, untested facility. Simulation was particularly useful due to the uncertain patient procedure and recovery times. We discuss the simulation process and show how it helped reduce patient waiting time while ensuring the clinic could meet its target patient volumes.

This paper attempts to integrate some of the recent research on several different aspects of scheduling production systems where early shipments are forbidden. In particular, this paper focuses on scheduling-rule development, order-release policies, due-date-setting policies, and the use of overtime on a selective basis. In addition, multiple performance criteria, ranging from net present value to average tardiness, are included. The various policies are examined using a simulation model of a hypothetical job shop and statistical significance is reported. The ‘best’ policies are given and shown to be dependent on the environmental conditions being modelled, such as utilization and tightness of due dates. The findings on overtime are that the net present value measure is not affected, but time-based measures are improved. In addition, delayed release is shown to be beneficial whenever job earliness is penalized.

Recent research has considered the use of economic objectives in production scheduling environments as opposed to the classic time-based measures prevalent in the literature and in practice. This paper presents a comparison of both types of objectives to determine how differently the objectives select scheduling methods and how much economic loss occurs when using time-based measures. A dynamic job shop is modeled using computer simulation to study the problem. Various utilization and due-date allowance settings are considered in the experimental design, along with a variety of time-based and economic objectives and scheduling methods. The results show that percent tardy was the most effective time-based objective in meeting economic goals. However, even with this measure substantial economic loss occurs under certain conditions. In general, the time-based measures do not lead to similar scheduling decisions, while the various economic methods tended to lead to at least “good” economic decisions.

In the past, performance in dynamic-scheduling environments was primarily measured in terms of time or physical shop characteristics. Objectives such as mean tardiness, flow time, and work-in-process inventory were commonly used. Today, there is increasing interest in the use of more advanced economic performance measures. These measures have the more comprehensive objective of maximizing ownership wealth by economically scheduling jobs and tasks. This study presents a large-scale experiment testing time-based and economic-based scheduling methods in a dynamic job shop. These methods are evaluated on their ability to maximize net present value (NPV). The study considers the just-in-time (JIT) delivery environment. The job shop is hypothetical, but is based on models of real production situations. Results show that the use of very detailed economic information in a sophisticated manner generally improves economic performance. Where due dates are easy to achieve, however, time-based scheduling methods are at least as good as those based on economics. Also, where utilization is high and due dates tight, early cost information in release and dispatch is detrimental to schedule value.

In a recent paper, Christy and Kanet (1990) examined the performance of several dispatching rules in a forbidden early shipment environment. Their primary performance measure was time-weighted inventory value. This study re-examines their results using a new model, with a net present value (NPV) performance measure. The results show that using NPV and inventory objectives lead to different scheduling decisions. Also examined are the reasons for relatively poor performance of operation-based due date rules.

Both practitioners and researchers in the field of Operations Management have suggested that shop scheduling should be an integral component in both the strategic and tactical plans for an organization's assets. This paper examines the use of an accepted measure of return on assets, net present value (NPV), in a simulated shop scheduling environment where early shipment of jobs before their due dates is forbidden. In addition, early shipment of raw materials to the shop is also forbidden. This shop environment is consistent with the prevalent practice in industry of accepting orders only on a just-in-time basis to reduce purchased parts inventories. The NPV measure provides a means of balancing a variety of performance criteria that have been treated as separate objectives previously, including work-in-process inventory, finished goods inventory, mean flow time and mean tardiness, while also providing a means of measuring monetarily the value of various shop scheduling approaches. The NPV performance of priority scheduling rules and order release policies is measured in this research through the simulation of a random job shop under a variety of environmental conditions. It is found in a comparison of priority rules that use time-based information with those that use job value information that the Critical Ratio rule provides higher average performance than the three other rules used in the study. However, in some situations that are consistent with JIT practice, value-based priority rules also perform well. The use of a mechanism for delaying the release of jobs to each work center in the shop provided higher average NPV when shop utilization was set at a low level of 80%, while immediate release of work upon its arrival to the shop provided superior performance at a higher shop utilization level of 94%. While JIT materials delivery and costing yields higher NPV, it did not alter the relative ranking of priority rule/release policy combinations. In addition, it was found that environmental factors, including average job length, average number of tasks per job and level of tardiness penalty, resulted in greater variations in NPV performance than the institution of a JIT raw materials delivery policy.

We report on the use of simulation modeling for redesigning phlebotomy and specimen collection centers (or patient service centers) at a medical diagnostic laboratory. Research was performed in an effort to improve patient service, in particular to reduce average waiting times as well as their variability. Discrete-event simulation modeling provided valuable input into new facility design decisions and showed the efficacy of pooling sources of variation, particularly patient demand and service times. Initial performance of the redesigned facilities was positive; however, dynamic feedback within the system of service centers eventually resulted in unanticipated performance problems. We show how a system dynamics model might have helped predict these implementation problems and suggest some ways to improve results.

Health-care consumers continue to be frustrated with long waits, especially when an appointment has been made. However, providers who book appointments are under increasing pressure to maximize utilization so that revenues will be increased and costs reduced. Thus, scheduling appointments involves opposing forces that are difficult to manage. This challenge is addressed in a rolling-horizon environment with fluctuating demand loads. These two issues have not been explored previously in the appointment-scheduling research. Two management policies are considered: overload rules (OLR) and rule delay (RD). The former considers different scheduling methods (overtime, double booking) when demand loads are high, and the rule delay policy considers when to implement the overload rules. These methods are explored for six different demand patterns/loads and evaluated with a variety of client and server-oriented measures. The results show that managers of appointment scheduling systems must carefully consider which measures are most important to them since the best choices of OLR and RD vary substantially by measure. Good choices also depend on the general type of client demand pattern. Thus, to consider the various tradeoffs between client and server measures a matrix is developed that outlines good choices for each scenario.