Daniel G. Shimshak

Stepwise selection of variables in data envelopment analysis: Procedures and managerial perspectives

Abstract One of the most important steps in the application of modeling using data envelopment analysis (DEA) is the choice of input and output variables. In this paper, we develop a formal procedure for a “stepwise” approach to variable selection that involves sequentially maximizing (or minimizing) the average change in the efficiencies as variables are added or dropped from the analysis. After developing the stepwise procedure, applications from classic DEA studies are presented and the new managerial insights gained from the stepwise procedure are discussed. We discuss how this easy to understand and intuitively sound method yields useful managerial results and assists in identifying DEA models that include variables with the largest impact on the DEA results.

Advances in physician profiling: the use of DEA

Abstract Insurers, health plans, and individual physicians in the United States are facing increasing pressures to reduce costs while maintaining quality. In this study, motivated by our work with a large managed care organization, we use readily available data from its claims database with data envelopment analysis (DEA) to examine physician practices within this organization. Currently the organization evaluates primary care physicians using a profile of 16 disparate ratios involving cost, utilization, and quality. We employed these same factors along with indicators of severity to develop a single, comprehensive measure of physician efficiency through DEA. DEA enabled us to identify a reference set of “best practice” physicians tailored to each inefficient physician. This paper presents a discussion of the selection of model inputs and outputs, the development of the DEA model using a “stepwise” approach, and a sensitivity analysis using superefficiency scores. The stepwise and superefficiency analyses required little extra computation and yielded useful insights into the reasons as to why certain physicians were found to be efficient. This paper demonstrates that DEA has advantages for physician profiling and usefully augments the current ratio-based reports.

Estimation of the evacuation time in an emergency situation in hospitals

In this paper, a prediction model is presented that estimates the evacuation time in an emergency situation for hospitals. The model is generic enough to be used in various hospital settings. This model can provide incident managers with estimates of the evacuation times of different types of patients and can offer support to the managers with their resource allocation decisions in emergency situations. The major advantage of the prediction model is that the computation time is very short and the model does not need a lengthy and costly design. The model was applied for several different evacuation scenarios and the results were compared with those of a simulation model which had already been designed for use by the hospital. The comparison shows that the prediction model can provide estimates of the evacuation time that are similar to the results found by using costly and time-consuming simulation models.

A Two‐Model Approach to Measuring Operating and Quality Efficiency with DEA

Abstract A serious problem in applying data envelopment analysis (DEA) to the study of service industries is the selection and application of quality output measures. Oftentimes, decision-making units (DMUs) are found to be efficient due to high operational output measures, even though they have low output quality. Furthermore, some of these DMUs are among the benchmark set for inefficient units. We present a Two-Model approach for including quality measures in DEA studies. This approach allows decision-makers to evaluate two models simultaneously, one measuring operational efficiency and the second measuring quality efficiency. This new method selects only DMUs that are efficient in both operational and quality measures to be members of the benchmark set. This paper demonstrates the Two-Model DEA approach using data from the nursing home industry. It can be applied to any situation in which either the DEA input or output measures can be represented as two distinct categories or objectives.

A priority queuing model of a hospital pharmacy unit

Abstract This study analyzes various design alternatives in determining the manpower requirements needed to efficiently run a hospitals pharmacy unit. Several queuing models are used in the analysis. Because of the complexity of the problem, various assumptions have to be made in modelling the pharmacys operations so that a solution using analytic methods can be found. This study was performed in order to provide the hospital administrator with a clear picture of what effect manpower will have on the waiting times of prescription orders under different conditions. In addition it gave quantitative data that supported current operating decisions. Three different operating procedures were evaluated in order to give a complete analysis of the prescription order process taking place in the pharmacy unit. These were a multiple server queuing model with no priorities, a priority discipline queuin model, and a priority discipline queuing model with preemptive service. Finally consideration was given to cost and waiting time objectives in order to determine the appropriate level of manpower.

Benchmarking nursing home performance at the state level

This paper reports the results of a state-level comparison of the performance of nursing homes. The results were obtained by applying data envelopment analysis (DEA) to data (obtained from OSCAR, the online survey, certification and reporting database [2004]) for all the skilled nursing facilities in each of the 50 states and the District of Columbia. DEA produces an efficiency score for each state that can serve as a single comprehensive measure of its overall performance. However, two DEA models were used in the analysis reported here so that each state could be given two efficiency scores, one for each of the two aspects of their performance – quality efficiency and operating efficiency. Eleven states were identified as being 100% efficient for both quality and operating efficiency of their nursing homes. For the remaining states (i.e. those having lower scores for either quality efficiency or operating efficiency or both), DEA results include, in addition to the efficiency score, performance goals and a benchmark set of other states that could be used as models of ‘best practice’.

An evaluation of alternative task allocations to two servers

The problem considered here is how to allocate n exponential tasks to two available servers. The servers may be placed in a series or parallel configuration, and arrivals are assumed to be Poisson. In general, this gives rise to a network of queues with Erlang service times. Alternative designs are examined using various approximation methods and simulation for several values of n and different utilization rates. If the workers must be placed in tandem, the rule of balancing the work of the two stations is found to be optimal for all cases considered. If it is possible to place the servers in a parallel configuration, each performing all n tasks, the average total waiting time can be substantially reduced.

Efficiency in Public Higher Education Systems: A Statewide Comparison Using Data Envelopment Analysis

As state funding for public higher education has declined, there is a rising demand for accountability. Past studies have relied on indicator ratios to look at the relationship between funding and performance measures. This approach has some inherent problems that make it difficult to identify inefficiencies. This chapter will study efficiency in state systems of higher education by applying data envelopment analysis (DEA). DEA methodology converts multiple variables into a single comprehensive measure of performance efficiency and has the ability to perform benchmarking for the purpose of establishing performance goals. The advantages of DEA modeling will be shown by comparing results with those from a recent study of higher education finance based on publicly available data. DEA is shown to be feasible and implementable for studying state systems of higher education, and provides useful information in identifying “best practice” state systems and guidance for improvement. The value of DEA modeling to state policy makers and education researchers is discussed.

Ranking Queue Systems Using Stochastic Dominance

AbstractThe appiicalion of stochastic ordering rules to waiting time distributions in single server queues is examined. Rules of this type can prove useful to the systems designer when analysing queueing systems. Systems can be compared in terms of average wailing (or system) time, in terms of the stochastic dominance of the waiting (or system) time distribution or in terms of the dominance of some other stochastic parameter characteristic of thesystem. Such rules are particularly important when one musi deal with non-linear waiting cost functions.

Analysis and redesign of a military orders processing system

Abstract A study was conducted which analyzed the army orders processing system at the Massachusetts Army National Guard. Timely processing of these orders was essential since they initiated active duty status. This status officially obligated a civilian into a military position. Orders processing behaved as a queuing system with batch arrivals and preemptive service. Given randomness of the arrival and service processes, insufficient clerk manpower caused the orders to be delayed in a queue. The analysis of this system was conducted using computer simulation modeling. The model was designed to schedule clerks so that total processing time was minimized and critical delays were avoided. The evaluation of different manpower alternative included varying the number and arrangement of the processing clerks. Ultimately, the entire orders processing system was redesigned. This article discusses the National Guards response to the results and recommendations determined through simulation modeling.