Journal of Operations Management 20 (2002) 389408

Technical note

The impact of strategic operations management decisions on community hospital performance

Ling X. Li a, , W.C. Benton b , G. Keong Leong c
b

Department of Information Systems and Decision Sciences, College of Business and Public Administration, Old Dominion University, Norfolk, VA 23529, USA Department of Management Sciences, Fisher College of Business, The Ohio State University, 2100 Neil Avenue, Columbus, OH 43210, USA c Department of Management, College of Business, University of Nevada Las Vegas, 4505 Maryland Parkway, Box 456009, Las Vegas, NV 89154-6009, USA Received 6 February 2001; accepted 14 November 2001

Abstract Over the past decade, 10% of community hospitals have closed. In this challenging time, our study presents hospital administrators with some valuable information that can help improve community hospitals performance. The purpose of this paper is to develop a strategic operations management model that links long-term service choices, intermediate operations decisions, and hospital performance given the structural constraints of location, size, and medical teaching status. Data collected from 151 community hospitals are used to test the model. The research identies strategic operations management decisions in the US community hospitals, shows their causal relationships, and identies their effects on hospital performance. Specically, we nd that intermediate infrastructural operations decisions affect a community hospitals cost, quality, and nancial performance after the structural decisions of location and size have set the stage. Our study also reveals that community hospitals have adopted new staff and demand management decisions in response to the market needs. 2002 Elsevier Science B.V. All rights reserved.
Keywords: Health service; Operations strategy; Infrastructural decisions; Community hospitals; Structural equation

1. Introduction Service is positioned at the center of economic activities in any society. As a matter of fact, no economy can function without the infrastructure that service organizations provide in the form of transportation, education, and health care (Fitzsimmons and Fitzsimmons, 2000). In recent years, we have witnessed a major evolution in the industrial nations, from being primarily manufacturing-based to primarily service-based. Therefore, knowing how to effec

Corresponding author.

tively manage a service organization has become a priority. Due to the increasing needs for better direction and management of service organizations, operations management researchers and practitioners have started to apply operations management concepts and techniques developed in manufacturing sectors to service industries (Heineke, 1995; Buler et al., 1996; Fitzsimmons and Fitzsimmons, 2000). One signicant segment of the service sector in the US is the health care industry. Health care service is a patient-oriented service that requires continuous interaction with customers. It utilizes facilities and equipment, and consumes a large volume of

0272-6963/02/$ see front matter 2002 Elsevier Science B.V. All rights reserved. PII: S 0 2 7 2 - 6 9 6 3 ( 0 2 ) 0 0 0 0 2 - 5

390

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

nursing care. Therefore, it becomes increasingly important to health care executives to understand what kind of facility, equipment, and workforce decisions are critical to achieve the commonly acknowledged goal of providing quality health service at a reasonable cost. The health care industry uses management models that were developed in manufacturing to improve its performance. In the late 1980s and early 1990s, health care professionals adapted quality control methods such as the Fishbone Diagram and quality control charts used by US manufacturers (Berwick et al., 1991). Recent health care research suggests that the health care industry can apply strategic operations management models developed in manufacturing to achieve better performance in the changing health care environment (Heineke, 1995; Roth and Van Dierdonck, 1995; Buler et al., 1996). Community hospitals account for 80% of US hospitals (AHA Survey of Hospital Data Base, 1995). A community hospital is a non-federal short-term general or other special hospital that is not a hospital unit of an institution (AHA Survey of Hospital Data Base, 1995). Until now research on operations strategy in community hospitals has not been well developed, especially when it relates operations strategy to the current health service condition. Most of the operations-oriented studies focus narrowly on issues of hospital cost containment, capacity planning, or personnel scheduling (Buler et al., 1996). For example, Roth discusses a strategic model that focuses on hospital technology management (Roth and Johnson, 1996), and another model that studies hospital resource management using the Material Requirements Planning concept (Roth and Van Dierdonck, 1995). Heineke (1995) applies the Hayes and Wheelwright strategic manufacturing model to health service, specically in the infrastructure management of obstetric departments. Through these innovative studies, the conceptual background for integrating strategic operations management to hospital administration is established. However, few examples can be cited where long-term structural decisions, intermediate operations decisions, and performance are included explicitly in the management of health service operation. Over the past decade, even as an aging population is consuming more health services, about

10% of US community hospitals have closed due to low occupancy rates and poor nancial performance (Anonymous, 2001). The existing community hospitals are trying to nd ways to survive, or even thrive, in the competitive health service market. Many strategic decisions, such as adding outpatient health service plans and partnering with physicians, have been used to improve market presence, utilization, and care condition (Edlin, 2001). At the same time, health care administrators have begun to realize that focusing on one decision area, such as structural decisions, is not enough. Instead, developing a congruent operations strategy is the key to improving a hospitals utilization and nancial performance (Buler et al., 1996). The purpose of our study is to develop a strategic operations management model that links long-term equipment and service choices, intermediate operations decisions, and performance given structural constraints for community hospitals. The expected contributions of this study include a better understanding of strategic operations management decisions in the US community hospitals, identifying the causal relationships among operations decisions, and recognizing their effects on hospital performance. The research is further expected to nd out the key strategic decisions that can help improve a community hospitals performance in the changing health service environment. There are two issues discussed in this study. First of all, given the structural constraints of location, size, and medical teaching status, what are the long-term service choices and intermediate infrastructural operations decisions developed by US community hospitals in response to the market needs in the current health service environment? Secondly, with various structural constraints, how do intermediate infrastructural operations decisions impact a hospitals cost and quality measures and nancial performance? The effects of intermediate decisions on performance are emphasized in this study because a previous study has suggested that health service, such as obstetric departments, relies upon a well-managed intermediate infrastructure decisions to deliver quality health service at a reasonable cost (Heineke, 1995). Several hypotheses related to the two research questions will be presented in the next section.

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

391

2. Conceptual model and hypotheses The hospital strategic operations management and performance model is presented in Fig. 1. This model is based on an extensive review of the literature (Buler et al., 1996; Hayes and Wheelwright, 1984; Heineke, 1995; Roth and Johnson, 1996; Li and Benton, 1996; Li, 1997) and direct discussions with health service experts (administrators of two hospitals in Ohio and one hospital in Michigan). The model explores the key determinants of a hospitals equipment and service choices and intermediate infrastructural operations decisions, and their effects on hospital performance with structural constraints. There is general agreement in the literature that operations management can make an important contribution to the success of an organization if operations decisions are made in support of the business strategy (Hayes and Wheelwright, 1984; Heineke, 1995; Skinner, 1986). These operations decisions are categorized as long-term structural decisions and intermediate infrastructural decisions and other productivity improvement programs such as quality management and continuous improvement (Hayes and

Wheelwright, 1984; Skinner, 1986). The detailed elements in the category of long-term and intermediate operations decisions may vary a little, but the major concept is consistent. Structural decision areas usually include location, facility, capacity, and product and equipment/process technology. These decisions are the bricks and mortar of an organization and are, therefore, regarded as having more signicant implications in a long-term time frame (Hayes and Wheelwright, 1984). Infrastructural operations decisions relate to workforce management, production planning and control, quality assurance, and organization design. These decisions are more tactical in nature, but their cumulative impact can be as important as that of structural decisions and can be as difcult and costly to change (Hayes and Wheelwright, 1984). 2.1. Long-term structural decisions Long term structural decisions are considered as the front end in a hospital resource planning system that denes the capacity resource needed to perform the various activities (Roth and Van Dierdonck, 1995). The long-term decisions discussed in this study

Fig. 1. Conceptual model.

392

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

are hospital location, bed size, medical education involvement, outpatient service expansion, and equipment/technology investment (Heineke, 1995; Roth and Van Dierdonck, 1995; Smith-Daniels et al., 1988). These ve long-term strategic decisions that hospitals make can be broadly grouped into two categories: facilities management (hospital location and bed size) and service choices (medical teaching status, outpatient service expansion, and equipment/technology investment). In this study hospital size, location and medical teaching status are treated as structural constraints because these decisions were made earlier and have become existing conditions now. 2.1.1. Long-term facility management decisions 2.1.1.1. Bed size. Bed size is an indication of the capacity of a hospital. The number of beds provides hospitals with economies of scale (Hancock et al., 1976; Nath and Sudharshan, 1994). Size has a direct relationship with service choice decisions. Large hospitals tend to be urban or medical education related hospitals and provide more service lines. By providing a variety of services, hospitals have increased their service volumes from different service lines (Nath and Sudharshan, 1994). During the past decade, cost containment pressures have forced hospitals to re-examine many of their facility management decisions. The number of inpatient admissions has declined and the average inpatient length of stay has fallen (Hospital Statistics, 1994; Smith-Daniels et al., 1988). A certain number of medical treatments and surgical procedures have been shifted from an inpatient basis to an outpatient basis. These changes have resulted in larger hospitals experiencing more demand variation than smaller ones. The decline in the hospital occupancy rate has led hospital facility management in a new direction. More specically, large hospitals are expanding their ambulatory service capacity to meet the outpatient service demand and forge partnerships with other hospitals to improve utilization (Comparative Performance, 1993). The related hypothesis here is H1a. The relative size of a hospital (number of beds) as a structural constraint has a positive effect on hospital service choices, including outpatient services and health care network.

The other noticeable issue relating to large hospitals is that larger hospitals are usually equipped more extensively to treat complex medical conditions; and therefore, they tend to invest more in equipment and technology since they have a larger customer base than smaller hospitals do. This, in turn, gives them a relatively comfortable demand volume to absorb the cost of the investment in equipment and technology. The equipment/technology related hypothesis is H1b. The relative size of hospitals (number of beds) as a structural constraint has a positive effect on hospital equipment/technology management decisions. 2.1.1.2. Location. Hospital location is another longterm facility decision that is likely to have an effect on service choices. The location of a hospital in a rural area as opposed to an urban area inuences its facility management choice (Smith-Daniels et al., 1988). For example, the distance to a health care center from a patients home may have a negative effect on hospital facility utilization. In recent years, hospitals in rural areas have been reinventing their operations strategies to match the competition. This is especially true in the area of continuous improvement (see Section 2.2.3 for more details) to promote job enrichment and skill exibility because demand for each unique labor skill is low and staff members have to be more versatile to overcome volume disadvantage (Henry, 1994; Hudson, 1995). Additionally, in recent years, rural hospitals have established partnerships with urban hospitals and physicians, and have created health service networks to strengthen rural residents access to high-quality health care services. The notion of mutual cooperation holds substantial appeal for rural communities and hospitals to meet the challenge of the current health service environment (Moscovice et al., 1995). The related hypotheses for location decisions are H1c. Urban and rural hospitals make different service choices because of their geographic location constraint. H1d. Urban and rural hospitals tailor their continuous improvement techniques based on their geographic location constraint.

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

393

2.1.2. Long-term service choices Service choices, in this study, include involvement in medical education, outpatient service and service network, and equipment/technology investment for technology-related services. All three choices can affect service complexity. 2.1.2.1. Medical education involvement. Medical education involvement is a long-term service decision that is likely to have an effect on hospital service planning and control decisions (Flood and Scott, 1987). Hospitals involved in medical education tend to be large. As the hospital occupancy ratio has decreased in recent years, these hospitals have to manage higher demand variability. Additionally, hospitals that have medical education programs usually treat more complicated cases that have higher demand variability than other hospitals. Therefore, patient admission, surgical scheduling focused on patient expected length of stay, and patient classication mix have become crucial decision variables for demand prediction and management (Grifth et al., 1976; Flood and Scott, 1987). It has been suggested that teaching hospitals adopt alternative capacity management decisions to mange demand variability (Smith-Daniels et al., 1988). The hypothesis related to hospital teaching status is H2a. The medical teaching status of a hospital as a structural constraint positively affects the hospitals demand management. 2.1.2.2. Outpatient service and health service network. A typical hospital is a multiple service provider. As such, a hospitals choice of services could affect equipment/technology management, workforce management, demand management, and quality management programs. In recent years, outpatient services are combined with inpatient services to provide a more integrated health service delivery system and expand the scope of services (Edlin, 2001). Additionally, the increased volume of outpatient surgery, to a certain degree, reects advanced health care technology that permits less invasive procedures and shorter patient recovery times. Moreover, the availability of various services will affect physicians selections for patient referral and inuence the patients and payers hospital preferences (Becker and Sloan, 1985; Nath and Sudharshan, 1994). However, a

wider range of services may require more extensive equipment/technology support, as well as more staff support and a well-developed demand management system. In addition to providing a broader range of services, the practice of strategic partnership has been adopted by the health care industry. Hospitals have started forging partnerships with managed care systems, physicians, local hospitals, and established networks to improve facility and equipment utilization so as to improve cost performance (Coddington and Moore, 1987; Hospital Statistics, 1994). This trend represents a restructured health care delivery system built around the integrated, community-based network of care (Burns et al., 1998; Edlin, 2001; Hospital Statistics, 1994; Smith-Daniels et al., 1988). The hypotheses related to service choices are H2b. The choice of providing outpatient services and participating in health service network will lead a hospital to improve demand management. H2c. The choice of providing outpatient services and participating in health service network will lead a hospital to improve workforce management. H2d. The choice of providing outpatient services and participating in health service network will lead a hospital to invest more in continuous improvement activities such as training, skill enhancement, and job enrichment. H2e. The choice of providing outpatient services and participating in health service network will lead a hospital to invest more in equipment and technology. 2.1.2.3. Equipment/technology decisions. In this study, we asked hospital administrators about the level of investment in clinical, ofce, and patient information system technology. Together, these technologies and equipment, with the support of knowledgeable staff, will enable hospitals to develop lasting competitive capabilities (Roth and Johnson, 1996). Since clinical processes generally revolve around service lines, hospitals usually make technology investment decisions based on the services they are offering or will offer. Therefore, equipment/technology decisions become an important component of service choices.

394

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

However, technology alone will not provide hospitals with a world-class status. It is sound infrastructure programs, such as readiness of people to use technology and quality assurance procedures, that give a hospital competitive capabilities (Roth and Johnson, 1996). The related hypothesis is H2f. Investment in equipment and technology will require the readiness of people through training and skill development (quality improvement decisions) to use that technology. 2.2. Intermediate operations decisions Hayes and Wheelwright (1984) suggested that infrastructural decisions consist of workforce, quality, production planning and control, and organization. These decisions are more tactical and are linked with specic operating aspects of an organization. We have adapted Hayes and Wheelwrights infrastructural concept to the health service setting in this paper and refer to infrastructural decisions as intermediate operations decisions, which include service planning and control, workforce management, and quality improvement decisions. In this study, we did not ask hospital administrators explicitly about organization design because this issue is embedded in both the hospitals long-term and intermediate operations decisions. For example, one of the long-term new organizational designs of US hospitals is to forge partnerships with physicians and other hospitals to provide integrated service to customers. 2.2.1. Demand management In hospitals, demand management resembles planning and control systems in a manufacturing rm. Demand management involves the guidelines for controlling demand and managing the ow of patients through the service system (Heineke, 1995). Demand for hospital services depends on many unpredictable factors, such as the time that services are needed and the type of professional skill that is required. Therefore, hospitals have to develop guidelines to manage the issues of hospital inpatient admission, inpatient and outpatient surgical schedules based on expected length of stay, and the mix of diagnosis related groups (DRG). These factors place constraints on hospital demand management.

Appropriate demand management decisions through planning and control can positively affect hospital cost performance. A responsive planning and control mechanism tends to better predict demand level and manage demand variability. The variables included are demand prediction and managing demand. The related hypothesis is H3a. Demand management will have a positive impact on hospital operational cost measures. 2.2.2. Workforce management The purpose of workforce management is to match the supply level with the demand level. Health service requires constant interaction with patients. This environment creates uncertainty in daily operations due to the unpredictable times of customer arrival and the demands of the patients. Therefore, an appropriate staff mix and effective staff scheduling procedure will help meet time varying demand while improving the cost of service (Siferd and Benton, 1994). In recent years, staff overage and shortage management has become an important cost control issue. Hospitals are increasing overtime usage when demand exceeds supply. Some hospitals send their nursing staff home with or without pay when there are more nurses than needed. Other hospitals are reconsidering the composition of their staff mix, such as the number of registered nurses and nurse aids, in order to stay under the cost containment pressure while not jeopardizing service quality. As a result of the current nursing supply shortage in community hospitals, patients are being diverted from understaffed emergency rooms to other relevant units (Egger, 2000; Johnson, 2001). In order to provide adequate services, community hospitals are actively seeking solutions to this issue. In this study, workforce management variables include staff composition, staff increase and reduction, and staff overage and shortage management. The related hypothesis is H3b. Better workforce management decisions will positively affect hospitals operational cost measures. 2.2.3. Continuous improvement Continuous improvement depends to a large extent on motivated and well-trained employees. Kaizen or continuous improvement is a Japanese concept of

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

395

continuously looking for methods to improve quality and operations. The belief is that any operation can be improved and employees are best positioned to identify and make the changes. Workers should be provided on-the-job training, staff skill development, and job enrichment (Deming, 1982). It is important that a hospital empowers its workforce through education and training. To retain health care service quality through continuous improvement while containing costs, hospitals are focusing on improving staff exibility, enhancing service innovation, and increasing staff training (Award Criteria, 1995). Further, the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) has set guidelines for the staff qualication and competence requirements for accreditation (Accreditation Manual, 1993). Evidently, a competent and knowledgeable staff is more likely to design and deliver services that conform to medical requirements and customer needs (Li, 1997). The variables included are staff training, staff competence, and job enrichment. The related hypothesis is H3c. Hospital continuous improvement activities will positively affect hospital quality performance. 2.3. Performance A hospitals nancial performance is crucial to a hospitals existence. It is reported that hospitals with above average performance tend to have congruent strategic operations management decisions (Buler et al., 1996). However, for the past decades, the quality advocators have recognized that undesirable quality performance tends to result when managers are evaluated solely on nancial measures (Crosby, 1979; Deming, 1982). In practice, separating the nancial and operational performance of a service provider will lead service administrators and staff members to divergent goals (Heineke, 1995). This study intends to link a hospitals cost and quality performance with its nancial performance. We suggest that improving cost and quality performance will lead to better nancial performance because good quality services tend to reduce waste during the service delivery process and prevent malpractice from occurring. Reductions in waste lead to superior cost performance, which in turn, improve a hospitals nancial performance.

We asked hospital administrators to compare their performance with that of their competitors. In general, a hospitals cost, quality, and nancial performance is measured by comparing the performance data of competing hospitals in the same area, by similar size facilities, or using other hospital characteristics (Comparative Performance of US Hospitals, 1993; Guide, 1994). The cost measurement areas of this study include holding down patient cost (Fetter et al., 1980; Flood and Scott, 1987), and attaining high labor productivity and capacity utilization. (Siferd and Benton, 1994). The quality measures include clinical quality (Donabedian, 1982) and customer satisfaction (Li and Benton, 1996). Customer satisfaction includes patient satisfaction, responding to patient requests, and responding to patient complaints. Financial measures included in this study are the hospitals market share, operating prot, and return on assets and investment. Many operations management studies such as Swamidass and Newell (1987), Vickery et al. (1993), and Ward et al. (1995), have used subjective performance measures. Ward et al. (1995) carried out a check on the self-reported performance measures in an earlier study and concluded that the informants characterization of changes in prots is veried by the public record. The related hypotheses are H4a. Better quality performance will positively affect hospital cost performance. H4b. Better cost performance will contribute to better hospital nancial performance.

3. Research methodology 3.1. Data The conceptual model presented in Fig. 1 was subjected to a structural equation model of hospital strategic operations management decisions (Fig. 2) and was tested using data from Hospital Service Management database established in 19941995. Questions used in this study are presented in Appendix A. The organizational entity studied was the community hospital because these hospitals offer services to the general public, and long-term operations strategy and intermediate operations decisions and performance

396

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

Fig. 2. Structural equation model.

could be readily studied at the hospital level. Data was collected using mail surveys. Hospital administrators and/or chief operating managers responded to the research instrument. Follow-up letters were sent and follow-up phone calls were made to encourage hospital administrators to participate in the study. The questionnaire was mailed to 492 hospitals from the 1994 American Hospital Association Guide; 165 responded. Fourteen cases were dropped due to incomplete responses, resulting in 151 hospitals for this study. Administrators of non-responding hospitals were contacted to explore the reasons for not responding to the survey. The excuses for not responding to the survey included the following: the chosen participant was no longer working for the hospital, mergers, too many surveys received by the hospitals, and the low priority given to answering questionnaires. The results from Chi-square tests indicate that there were no signicant differences between respondents and non-respondents in terms of hospital size, location, and medical education involvement.

3.2. Scales and constructs An assessment of internal consistency and reliability of the measurement scale was conducted prior to the evaluation of the conceptual relationships in the proposed model. Specically, Cronbachs alpha, a measure for testing the internal consistency or reliability of a set of two or more scale indicators (Cronbach, 1951), was computed for each set of measurements (see Appendix A). All of the Cronbachs alpha values met the minimum criterion alpha value of 0.60 as suggested by Nunnally (1978), except outpatient service (0.55), health service network (0.58), and demand management (0.55). For exploratory work, the value of Cronbachs coefcient alpha of 0.50 to 0.60 is considered adequate (Nunnally, 1978). Content validity, which species that the research instrument reects the domain of the research area, was established through extensive interviews of practicing health service administrators from several functional areas of the three hospitals to identify those strategic operations management issues

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

397

that are viewed as important by managers. Additionally, each research construct in the conceptual model was validated through a comprehensive literature review and a pilot-test of the survey instrument at three hospitals. The survey instrument was revised according to the suggestions of the practicing managers and the results from the pilot study. This process validated the survey items in an objective manner. A seven point Likert scale was used to collect data. Most statements had response categories ranging from no emphasis (1) to extreme emphasis (7). Several statements were coded with seldom (1) to always (7). The questions on equipment/technology investment were coded from no investment (1) to heavy investment (7). The questions on performance ranged from signicantly low (1) to signicantly high (7) (see Appendix A). Several questions on the same concept were asked to get a better approximation of the scale whenever reasonable. In the operations management literature, when each of the scales is unidimensional, a single set of factor scores is used to indicate each scale (Ward et al., 1995). A factor-based scale is obtained by averaging the response to the items based on factor loadings. The result is a set of factor-based scores for each hospital for each scale. The covariance structural model for hospital strategic operations management decisions and performance is illustrated in Fig. 2. The three rectangular boxes, medical education involvement, hospital location, size at the left-hand side, are manifest (observed) variables. Outpatient service and service network, equipment/technology investment, demand management, workforce management, and continuous improvement activities are latent variables. The performance criteria, including cost, quality, and nancial performance are shown on the right hand side of the gure. 3.3. Covariance structure model We used covariance structure analysis to test the t of the representative model and to estimate the direct and indirect effects of long-term structural decisions on the choices of operations strategies and performance. The covariance structure model is a multivariate analysis methodology for empirically examining

sets of relationships represented in the form of linear causal models (Bollen, 1989; Joreskog and Sorbom, 1989). It provides researchers with a comprehensive means for assessing and modifying theoretical models. Therefore, it is an appropriate methodology for empirically testing the conceptual model proposed in this study that species causal relationships between a number of observed and latent variables. A two-step approach was taken to estimate the covariance structure model (Anderson and Gerbing, 1988). This approach suggests that the model testing task be considered as two conceptually distinct models: a measurement model that species the relations of the observed measures to their posited underlying constructs, with the constructs allowed to inter-correlate freely; and the structural model then species the causal relations of the constructs to one another, as posited by theory. According to Anderson and Gerbing (1988, p. 411), there is much to gain in theory validating, testing, and the assessment of construct validity from separate estimation of the measurement model prior to the simultaneous estimation of the measurement and structural sub-models. We, therefore, followed the two-step procedure recommended by Anderson and Gerbin. The SAS Systems CALIS procedure was used to analyze the data (SAS Institute Inc., 1990).

4. Findings In this section, we report the general ndings related to the measurement and structural models. We also present the ndings associated with the two research questions. The rst question is: given the structural constraints of location, size, and medical teaching status, what are the long-term service choices and intermediate infrastructural decisions developed by US community hospitals in response to the market needs in the current health service environment? This question is answered using ndings for Hypotheses 1 and 2. The second question is: with various structural constraints, how do infrastructural (intermediate) operations decisions impact a hospitals cost and quality measures and nancial performance? This question is answered using ndings from Hypotheses 3 and 4. A discussion of the more important ndings is provided in Section 5.

398

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

4.1. General ndings 4.1.1. Measurement model A measurement model describes the nature of the relationship between a number of latent variables, and the manifest indicator variables that measure those latent variables. The model studied in this research consisted of ve latent variables: (1) outpatient service and network, (2) equipment and technology investment, (3) demand management, (4) workforce management, and (5) continuous improvement activities. The goodness of t indices for the specied measurement model are reported in Table 1. This table shows that the Chi-square value for the measurement model is insignicant: 2 is 115 (d.f . = 118) and p = 0.54. An insignicant Chi-square value is more desirable and indicates that the conceptual model ts the empirical data (Joreskog and Sorbom, 1989). Overall goodness of t indices for the model are acceptable. The value for the goodness of t index (GFI) is 0.93 and adjusted goodness of t index (AGFI) 0.89, both exceeding the suggested value of 0.9 for GFI and 0.8 for AGFI (Bentler, 1990). The comparative t index (CFI) is 1.0 and non-normed t index (NNFI) is 1.0 as well. Both are in excess of the suggested value of 0.9 (Bentler and Bonett, 1980; Bentler, 1990). The normed t index (NFI) is 0.84 (0.60 has been suggested as an ad hoc criterion, Netemeyer et al., 1990). 4.1.1.1. Outpatient service and network. The two indicators of the outpatient service and network construct are outpatient service (x4 ) and health care network (x5 ). These two indicators have signicant standardized factor loading at p < 0.01 (Table 1). This nding shows that the service scope of many hospitals has expanded from traditional inpatient acute care to include outpatient and service networking. 4.1.1.2. Equipment/technology investment. The key indicators associated with the equipment/technology investment construct in the changing health care environment are investment in ofce technology (x17 ), clinical technology (x18 ), and patient medical information systems (x19 ). All three indicators have a standardized path coefcient signicant at a p-value of p < 0.01 (Table 1). The ndings support our suggested

model that hospital delivery systems are ripe for investing in advanced technology. 4.1.1.3. Demand management. The key indicators associated with the demand management construct are demand prediction (x6 ) and managing demand (x7 ). Both indicators have a standardized path coefcient that is signicant at a p-value of p < 0.01 (Table 1). 4.1.1.4. Workforce management. The key indicators associated with the workforce management construct include staff composition adjustment (x9 ), staff shortage management (x11 ), and excess staff capacity management using non-pay related measures (x12 ). All three indicators have a standardized path coefcient signicant at a p-value of p < 0.01 (Table 1). Three staff management indicators were not statistically signicant, and therefore, were dropped during the measurement model rening process. They are staff reduction (x8 ), staff augmentation (x10 ), and managing excess staff capacity using pay-related measures (x13 ). 4.1.1.5. Continuous improvement. Three key indicators associated with the continuous improvement construct in the changing health care environment have been identied. They are staff training (x14 ), staff competence (x15 ), and job enrichment (x16 ). All three indicators have signicant standardized path coefcients at a p-value of p < 0.01 (Table 1, Fig. 2). 4.1.2. Structural model The structural model determines the causal relationship among the constructs. Fit indices for the structural model are presented in Table 1. It can be seen that the t indices for GFI, AGFI, CFI, and NNFI are all above 0.9 and the 2 is 119 (d.f . = 136) with a p-value of 0.84. The standardized path coefcients for the study are also reported in Table 1. Combining the ndings of t indices obtained from the measurement model and structural model, we can see that the sample data support our conceptual model. 4.2. Findings related to Hypothesis 1 We hypothesized that hospital bed-size is an important decision variable in service choice decisions (H1a). As hospital size increases, service choices (such as outpatient service and technology investment) will

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408 Table 1 Model tting statistics Model goodness of t statistics and degrees of freedom p-value Goodness of t index (GFI) Adjusted (AGFI) Comparative t index (CFI) Normed t index (NFI) Non-normed t index (NNFI) Constructs & indicators from measurement model (SERV) (y1 ) outpatient service (x4 ) Health service networking (x5 ) (EQUP) (y2 ) ofce technology (x17 ) Clinical technology (x18 ) Patient information system (x19 ) (DEM) (y3 ) demand forecast (x6 ) Managing demand (x7 ) (WKFRC) (y4 ) dropped during model rening (x8 ) Staff composition (x9 ) Dropped during model rening (x10 ) Staff shortage management (x11 ) Staff overage management (x12 ) Dropped during model rening (x13 ) (CI) (y5 ) training (x14 ) Competence (x15 ) Job enrichment (x16 ) Paths from the covariance structure model Location outpatient service & network Location continuous improvement Size outpatient service & network Size Equipment/technology investment Medical education demand management Outpatient service & network demand management Outpatient service & network workforce management Outpatient service & network continuous improvement Outpatient service & network Equipment/technology investment Equipment/technology investment continuous improvement Demand management cost Workforce management cost (not signicant, dropped during model rening) Continuous improvement quality performance Quality performance cost Cost nancial performance 2 Measurement model 115, d.f. = 118 0.54 0.93 0.89 1.00 0.84 1.00 Standardized loading 0.64 0.51 0.84 0.66 0.76 0.51 0.55 0.40 0.42 0.53 0.67 0.68 0.71 Standardized path coefcient 13 53 12 22 31 31 41 51 21 = 0.14 = 0.30 = 0.37 = 0.20 = 0.25 = 0.81 = 0.83 = 0.51 = 0.47 Covariance structure model 119, d.f. = 136 0.84 0.93 0.90 1.00 0.83 1.00 t-value 6.24 5.33 11.32 8.33 10.05 4.52 4.73 4.00 4.17 5.13 7.91 8.05 8.40 t-value 1.28 3.42 1.70 4.39 2.16 11.86 6.30 5.11 5.62 1.41 4.64 1.84 4.41 5.85

399

p-value <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 p-value <0.10 <0.01 <0.05 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.10 <0.01 <0.05 <0.01 <0.01

52 = 0.20 63 = 0.42 75 = 0.17 67 = 0.32 86 = 0.43

increase as shown from our result (12 = 0.37, Fig. 2, standardized path coefcient). Given a xed number of beds each day in a hospital, occupancy rates of <100% would represent wasted capacity resources. As the occupancy ratio decreases nationwide, larger

hospitals may have more idle facility capacity than smaller ones, or larger hospitals may be able to handle higher demand variability. Having observed this phenomenon from the data, larger hospitals are more interested in expanding outpatient services, forging

400

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

partnerships with physicians and other hospitals, and seeking effective demand management decisions to improve utilization and achieve better cost performance (y6 in Appendix A) (H1a). Additionally, we nd that larger hospitals tend to invest more in equipment and technology for providing technology related services as compared to smaller hospitals (22 = 0.20) (H1b). One of the links of operations decisions is characterized by location and service choice. We hypothesized that hospital location as a structural constraint has a causal relationship with a hospitals service choice decisions, which could inuence intermediate operations decisions, and cost performance (H1c). As expected, hospital location has a signicant and positive impact on service choice decisions (13 = 0.14, standardized path coefcient, Fig. 2). This implies that urban community hospitals tend to expand their service scope from primarily inpatient service to a combination of inpatient, outpatient surgery, and clinic service (H1c). We hypothesized that urban and rural hospitals consider the nature of their geographic location while addressing continuous improvement issues (H1d). Our results indicate that there is a difference between rural and urban hospitals in their focus on staff training and skill development (53 = 0.30, Fig. 2; the negative sign is due to the 01 coding method, 0: rural and 1: urban). The rural hospitals tend to place more emphasis on staff development than their counterparts in the urban areas. Over the past decade, the decline in the number of community hospitals in rural areas has been sharper and more persistent than the decline of urban community hospitals (Hospital Statistics, 19941995). In order to survive the competition, rural community hospitals are developing effective continuous improvement activities (Henry, 1994; Hudson, 1995) such as providing staff members with on-the-job training, improving staff members skills, and implementing job sharing. Through these continuous improvement activities, employees are more versatile in handling a variety of tasks with a low demand volume of each service line. This nding supports Hypothesis H1d. 4.3. Findings related to Hypothesis 2 Hospital medical teaching status as a structural constraint is hypothesized to affect hospital demand

management (H2a). Medical teaching status shows a statistically signicant relationship with demand management as shown in Fig. 2 (31 = 0.25, t = 2.16). A possible explanation is that hospitals offering medical education programs tend to be large and treat more complicated cases. This service nature leads to higher demand variability which requires a better demand management system. Demand management characterizes one of the causal relationships of service choices and intermediate operation decisions. We hypothesized that a hospitals choice of providing outpatient services and participating in health service network would have a causal relationship with hospital demand management, which would inuence cost performance (H2b). As expected, we found that service choice has a signicant and positive impact on equipment/technology decisions (31 = 0.81, standardized path coefcient, Fig. 2). One of the causal relationships of service choices and intermediate operation decisions is characterized by workforce management decisions. We hypothesized that a hospitals choice of providing outpatient services and participating in health service network would have a causal relationship with workforce management decisions, which would inuence cost performance (H2c). As expected, service choice has a signicant and positive impact on workforce management decisions (41 = 0.83, standardized path coefcient, Fig. 2). An additional causal relationship of service choices and intermediate operation decisions is characterized by continuous improvement activities. We hypothesized that a hospitals choice of providing outpatient services and participating in health service network would have a causal relationship with continuous improvement activities, which would inuence cost performance (H2d). As expected, our result shows that service choice has a signicant and positive impact on continuous improvement activities (y7 in Appendix A) (51 = 0.51, standardized path coefcient, Fig. 2). One of the concerns of the relationship between a hospitals long-term structural decisions and intermediate operations choices is whether hospital service choice will have a causal effect on that hospitals equipment/technology investment decisions, which in turn inuence staff skill decisions and quality

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

401

performance. Our study shows that each service may impose unique requirement(s) on equipment decisions, which, in turn, affect quality decisions. This nding suggests that equipment and technology needed to support service choice decisions must be assessed carefully when a hospital plans to expand its services because long-term choices affect intermediate decisions. As expected, service choice has a signicant and positive impact on equipment/technology decisions (21 = 0.47, standardized path coefcient, Fig. 2). This nding supports hypothesis H2e. Technology investment decisions also require sound supporting infrastructure in order to gain superior competence over the competition. Infrastructure programs include the readiness of staff to use technology, quality assurance procedures, and system integration mechanisms. These programs should be aligned with the service and technology choices a hospital has made (52 = 0.20, standardized path coefcient, Fig. 2). This nding supports H2f. 4.4. Findings related to Hypothesis 3 We hypothesized that better demand management, workforce decisions, and equipment/technology decisions would have a direct effect on hospital cost performance (H3a). The results of the study suggest that better demand management will improve cost performance as shown by the standardized coefcient of 0.42 for 63 (Fig. 2). Hospitals that put more effort into demand management are better in improving capacity utilization, holding down average patient cost, and achieving higher labor productivity as compared to their competitors. The path of workforce management decisions and cost control efforts was found to be insignicant and dropped from the model tting process. As such, workforce management decisions do not directly contribute to cost performance (H3b). Continuous improvement (y7 in Appendix A) has a positive effect on quality performance (75 = 0.17, Fig. 2). This result shows that hospitals would perform better than their competitors by enhancing clinical quality, improving customer satisfaction, and responding to patient requests and complaints by increasing their staff exibility and training efforts (H3c).

4.5. Findings related to Hypothesis 4 Our results show that hospital nancial performance (y8 in Appendix A) is indirectly affected by quality measures (67 = 0.32, Fig. 2) (H4a) and directly affected by cost measures (86 = 0.43, Fig. 2) (H4b). A congruent strategic operations management model will positively contribute to hospital performance. 4.6. Summary of ndings The major contribution of this study is to develop a congruent model of long-term service choices, intermediate infrastructural decisions and performance given the structural constraints of location, size, and medical teaching status for community hospitals in the current health service environment. The model has been tested using 15 hypotheses and data from 151 US community hospitals. We have identied the key operations decisions made by US community hospitals in response to the market needs in the current health service environment. The results of our study show that given the structural constraint of size, larger hospitals are more interested in expanding outpatient services, forging partnerships with physicians and other hospitals, and seeking effective demand management decisions to improve utilization and achieve better cost performance. Additionally, we nd that larger hospitals tend to invest more in equipment and technology for providing technology related services as compared to smaller hospitals. Small rural hospitals, on the contrary, tend to place more emphasis on staff development than their counterparts in the urban areas. Given the structural constraints, we nd that intermediate infrastructural decisions affect a community hospitals cost, quality, and nancial performance in a challenging time when 10% of community hospitals have closed in the past ten years. The nding is consistent with the report from the manufacturing industry that low performing rms should rst place more emphasis on infrastrucutral operations decisions than on structural decisions. After shaping a sound infrastrucutral foundation, these rms may choose appropriate structural decisions and benet from the long-term decisions (Clark, 1996; Grant et al., 1991). Our study on the causal relationship between intermediate infrastructural decisions and performance in the health

402

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

care industry is an important antecedent to substantive structural decisions. Hospital administrators are provided with an idea of what operations decisions should be developed rst in order to manage declining occupancy rates, shortages of nurses, and poor nancial performances. Additionally, evidence from our study shows that hospital administrators have responded effectively to the current market demand through developing integrated long-term and intermediate operations decisions, such as linking service choices (long-term) with demand management (intermediate), linking equipment and technology investment (long-term) with human resource investment decisions (intermediate), and linking inpatient acute service with outpatient surgery and home care services (long-term). These integrated decisions directly and indirectly affect hospitals cost, quality, and nancial performance.

5. Discussion The covariance structural model shown in Fig. 2 provides us with some valuable insights into the role of long-term service choices and intermediate operations decisions on hospital performance given the structural constraints of location, size, and medical teaching status. Through the empirical analysis of the role of strategic operations management decisions on performance, we see that health care managers have responded effectively to the current economic and political mandates for decreasing costs and increasing health care quality. Many hospitals have sought to improve their overall competitiveness by eliminating badly fragmented and wasteful health service delivery processes. Due to added structural constraints, US community hospitals have emphasized demand management to better meet varying service requirements, have created a broad service scope that includes inpatient acute care, outpatient surgery, and home health services, and have maximized equipment and technology leverage to support a broader service spectrum. These integrated decisions directly and indirectly contribute to the more desirable cost and quality performance, which leading to a better nancial performance. Additionally, our study reveals the causal relationship between service choices and intermediate

decisions such as demand management and workforce management. Given the constraints of location and size, hospitals have forged partnerships with physicians and other hospitals. Most physicians are not hospital employees, but are granted privileges to practice health service in hospitals. What physicians do or do not do, and what they say or do not say can affect all hospital decisions, including the decisions to diversify into new services and new markets (Shortell, 1989). Therefore, forging partnerships with physicians can enlarge service scope and ensure a service volume for a hospital. This volume will affect intermediate decisions such as demand management and workforce management, and will indirectly affect cost performance. Our study reveals that one of the obvious differences in previous hospital research and current hospital practice in managing demand is that previous research tends to focus on a reactive approach to manage demand through internal improvement of facility utilization and better scheduling policies (Hancock et al., 1976). The current practice takes a more proactive approach through predicting demand requirements and scheduling the demand in accordance to the supply level. The demand management choice made by hospital administrators in recent years is to incorporate health service cost control, utilization control, and the coordination of care across institutional boundaries into demand management. For example, our study shows that when there are more patients than available beds, patients may be served at some other units to provide on time service and improve utilization. Early discharges may be solicited when it is legally and medically appropriate. Some health care publications report that hospitals have cut down the size of their staff under the pressure of cost containment (Comparative performance of US hospitals, 1993). However, our results indicate that staff reduction and addition are not signicant correlates of the workforce constructs. This implies that hospitals once did, but are no longer solely depending upon reductions in or additions to the number of workers on the payroll to meet demand variability, though they may still do so. Rather, hospitals are emphasizing an appropriate staff composition, using exible work time such as overtime, and exible work force such as a hospital nurse pool to manage demand uctuation. These exible workforce management methods

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

403

t better with the time varying demand that is so characteristic of a service organization. As indicated by the results of our study, gaining new skills is a key competence of health service workers. For the health service professional, continuous improvement through training and education is expected to match the requirement of advancing science and technology. Health service professionals are responsible for the efciency and effectiveness of care, and are beginning to use e-business communication modes and quality improvement methods. The health service providers also interface with other participants of the health care system, including community and specialized care services. Evidence from our study shows that hospitals are concentrating on professional qualications to meet patient requirements for care rather than on numbers in a stafng matrix. On-the-job training and competence development are necessary components of quality management policies. Our study indicates that given the structural constraint of location, hospitals are forming health service partnership based networks. This kind of network aims to provide the public with access to a broad spectrum of service and seamless coordinated care. Tomorrows health care professionals need to learn to grow in a new networked world. Compared to the more traditional health service organization of the past, the networked health service providers follow different rules and require different capabilities that need to be developed. The role of technology in health care cannot be over emphasized (Xu, 1996). Evidence from our analysis shows that technology is a means of inuencing quality assurance measures. Clinical technology and patient medical information systems remain the factors that allow hospitals to support changes (Xu, 1994a,b). Ongoing introduction of new technology has denitely changed the health care providers perspective on continuing education. Learning new competencies is now an accepted norm to keep up with the changing environment.

6. Future research, limitations and conclusion There are several ways in which this study can be extended. An extensive eld study could be conducted

to investigate the best practices for the various regions of the US hospitals could then determine how they measure up against the best practices in their region. Additionally, the study suggests that skill enhancement activities such as staff training, job enlargement, and staff competence are the most important staff management variables. Future research could consider developing a more generalizable workforce measurement scale. This scale could then be used to screen current and prospective employees. Additional areas for study include staff development plans that incorporate the concept of life-long learning to better facilitate the need of becoming a learning organization (Squires, 2001). Another important workforce management issue is the continuing nursing shortage (Johnson, 2001). An integrated solution should be sought from educators, practitioners, policymakers, payers, and hospitals administrators to recruit and retain nurses. Finally, the optimal mix of outpatient services versus inpatient services for a large urban hospital should be investigated. Our research reports the result of the effects of managerial decisions on performance in a single period. We believe that the outcome of the rst periods decisions will have a reverse effect on both structural and infrastructural decisions and performance of the next period. We think this kind of reverse effect will be best addressed in a longitudinal study. Therefore, future research could investigate the impact of managerial decisions on performance in a multi-period study to better understand the subsequent effect of operations strategies on performance. Although this study has identied some valuable operations decisions that can help improve a community hospitals performance, and has established the link between long-term equipment and service choices, and intermediate operations decisions and performance given structural constraints in hospitals, there are some limitations that should be taken into consideration for future research. One limitation is performance measures. We used self-reported performance information in our study because actual performance data are not commonly available. For example, starting from 1995, the annual survey of hospitals conducted by the American Hospital

404

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

Association (AHA) only reports hospitals operating expenses. Revenue and other nancial data are not reported due to business condentiality concerns. Future research may consider using the case study method to analyze some of the best practice hospitals in the US using actual performance data, or to analyze hospitals operating efciency using AHAs hospital expense information to compare hospital performance. Business strategy and the environment have an impact on hospital operations strategy. Future research can integrate business strategy and environment into the model to provide a more holistic explanation of the operational issues facing different types of hospitals such as VA hospitals, community hospitals, and research hospitals.

In this paper, we have developed a strategic operations management model that links long-term equipment and service choices, intermediate operations decisions, and hospital performance, identied the causal relationships among operations decisions, and recognized their effects on hospital performance. Hospitals can use the results of our study to decide which capability to develop in order to compete in the health care industry and achieve better performance.

Acknowledgements The authors want to thank the Editor for his insightful criticisms, encouragement, and valuable suggestions.

Appendix A. Scales and constructs

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

405

406

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408

407

References
Accreditation manual for hospitals, 1993. The Joint Commission on Accreditation of Healthcare Organizations, Vol. 1. Oakbrook, Terrace, IL. AHA Survey of Hospital Data Base, American Hospital Association, 1995. Anderson, J.C., Gerbing, D.W., 1988. Structural equation modeling in practice: a review and recommended two-step approach. Psychological Bulletin 103, 411423. Anonymous, 2001. Business: Reconstructive surgery. Economist 358 (8206), 65. Award Criteria, 1995. Malcolm Baldrige National Quality Award, Gaithersburg, United States Department of Commerce, MD. Becker, E.R., Sloan, F.A., 1985. Hospital ownership and performance. Economic Inquiry 23, 2136. Bentler, P.M., 1990. Comparative t indexes in structural models. Psychological Bulletin 107 (2), 238246. Bentler, P.M., Bonett, D.G., 1980. Signicance tests and goodness-of-t in the analysis of covariance structures. Psychological Bulletin 88, 588606.

Berwick, D.M., Godfrey, A.B., Roessner, J., 1991. Curing Health Care, Jossey-Bass, San Francisco, CA. Bollen, K.A., 1989. Structural Equations with Latent Variables. Wiley, New York. Buler, T.W., Leong, G.K., Everett, L.N., 1996. The operations management role in hospital strategic planning. Journal of Operations Management 14, 137156. Burns, L., Morrisey, M., Alexander, J., Johnson, V., 1998. Managed care and processes to integrate physicians/hospitals. Health Care Management Review 23 (4), 7080. Clark, K.B., 1996. Competing through manufacturing and the manufacturing paradigm: is manufacturing strategy pass? Production and Operations Management 5, 4258. Coddington, D.C., Moore, K.D., 1987. Market-driven strategies in health care. Jossey-Bass, San Francisco, CA. Comparative Performance of US Hospitals: The Sourcebook, 1993. HCIA Inc. & Delotte & Touche, Baltimore, MD. Cronbach, L.J., 1951. Coefcient alpha and the internal structure of tests. Psychometrika 16, 297334. Crosby, P.B., 1979. Quality is Free. New American Library, New York.

408

L.X. Li et al. / Journal of Operations Management 20 (2002) 389408 Moscovice, I., Christianson, J., Johnson, J., Kralewski, J., Manning, W., 1995. Building Rural Hospital Networks. Health Administration Press, Ann Arbor, Michigan. Nath, D., Sudharshan, D., 1994. Measuring strategy coherence through patterns of strategic choices. Strategic Management Journal 15 (1), 4351. Netemeyer, R.G., Johnsoton, M.W., Burton, S., 1990. Analysis of role conict and role ambiguity in a structural equations framework. Journal of Applied Psychology 75, 148157. Nunnally, J.C., 1978. Psycometric Methods, McGraw-Hill, New York. Roth, A.V., Van Dierdonck, R., 1995. Hospital resource planning: concepts, feasibility, and framework. Production and Operations Management 4 (1), 229. Roth, A.V., Johnson, S.P., 1996. Strategic deployment of technology in hospitals: evidence for reengineering. Managing Technology in Health Care, 189214. SAS/STAT Users Guide, 1990. Version 6, 4th Edition. SAS Institute, Cary, NC. Shortell, S.M., 1989. The keys to successful diversication: lessons from leading hospital systems. Hospital and Health Science Administration 34 (4), 471489. Siferd, S., Benton, W.C., 1994. A decision model for shift scheduling of nurses. The European Journal of Operational research 74, 519527. Skinner, W., 1986. The productive paradox. Harvard Business Review, 5559. Smith-Daniels, V.L., Schweikhart, S.B., Smith-Daniels, D.E., 1988. Capacity management in health care services: review and future research directions. Decision Sciences 19, 889919. Squires, A., 2001. Sink-or-swim tactics? Nursing Management 32 (3), 33 and 35. Swamidass, P.M., Newell, W.T., 1987. Manufacturing strategy, environmental uncertainty and performance. Management Science 33 (4), 509524. Vickery, S.K., Droge, C., Markland, R.E., 1993. Production competence and business strategy: do they affect business performance. Decision Sciences 24 (2), 435455. Ward, P., Leong, G.K., Sum, C., 1995. Business environment, operations strategy, and performance: an empirical study of Singapore manufacturing. Journal of Operations Management 13, 99115. Xu, L., 1994a. Developing a case-based knowledge system for AIDS prevention. Expert Systems 11, 237244. Xu, L., 1994b. A decision support system for AIDS intervention and prevention. International Journal of Bio-Medical Computing 36, 281291. Xu, L., 1996. An integrated rule- and case-based approach to AIDS initial assessment. Developing a case-based knowledge system for AIDS prevention. International Journal of Bio-Medical Computing 40, 197207.

Deming, E.W., 1982. Improvement of quality and productivity through action by management. National Productivity Review 1 (1), 1222. Donabedian, A., 1982. The Criteria and Standard of Quality. Health Administration Press, Ann Arbor, MI. Edlin, M., 2001. For the modern IDS, less is often more. Managed Healthcare Executives 11 (3), 2630. Egger, E., 2000. Nurse shortage worse than you think, but sensitivity may help retain nurses. Health Care Strategic Management 18 (5), 1617. Fetter, R.B., Freeman, S.Y., Averill, R.F., Thompson, J.D., 1980. Case Mix Denition by Diagnosis-Related Groups. Medical Care, 18 (Suppl.). Fitzsimmons, J.A., Fitzsimmons, M.J., 2000. Service Management: Operations, Strategy, and Information Technology. McGraw-Hill, Boston. Flood, A.B., Scott, W.R., 1987. Hospital Structure and Performance. The John Hopkins University Press, Baltimore, MD. Grant, RM, Krishnan, R., Shani, A.B., Baer, R., 1991. A Appropriate Manufacturing Technology: A Strategic Approach. Sloan Management Review, Fall, pp. 4354. Grifth, J.R., Hancock, W.M., Munson, F.C., 1976. Cost control in hospitals. Health Administration Press, Ann Arbor, MI. Guide, The 1994 AHA, 1994. American Hospital Association, Chicago, IL. Hancock, W.M., Warner, D.M., Heda, S. Fuchs, P., 1976. Admission scheduling and control systems. In: Grifth, J.R., Hancock, W.M., Munson, F.C. (Eds.), Cost Control in Hospitals, Health Administration Press, Ann Arbor, MI. Hayes, R.H., Wheelwright, S.C., 1984. Restoring Our Competitive Edge: Competing Through Manufacturing. Wiley, New York. Heineke, J., 1995. Strategic operations management decisions and professional performance in US HMOs. Journal of Operations Management 13, 255272. Henry, W.F., 1994. Reinventing rural hospitals. Trustee 47 (8), 2223. Hospital Statistics, 19941995 Edition, 1994. American Hospital Association, Chicago, IL. Hudson, T., 1995. Rural priorities. Hospitals and Health Networks 69 (4), 4044. Johnson, D., 2001. Retention may be key to nurse shortage. Health Care Strategic Management 19 (3), 23. Joreskog, K.G., Sorbom, D., 1989. LISREL 7: A Guide to the Program and Applications, 2nd Edition. SPSS Inc., Chicago. Li, L.X., 1997. Relationships between determinants of hospital quality management and service quality performancepath analytic model. OMEGA 25 (5), 535545. Li, L.X., Benton, W.C., 1996. Performance measurement criteria in health care organizations. The European Journal of Operational Research 93, 449468.