Determinants of hospital closure in South Korea: Use of a hierarchical generalized linear model☆
Introduction
Closure of a hospital directly affects the availability of health care in many neighborhoods (Bindman & Keane, 1990; Rosenbach & Dayhoff, 1995; Samuels, Cunningham, & Choi, 1991), creates unemployment for health care workers, and thus is likely to affect the health of residents and employees (Jackson & Whyte, 1998). The locale's economy and related industry can be also affected by the closure of a hospital (Probst, Samuels, Hussey, Berry, & Ricketts, 1999), even though the closure may improve the operating efficiency of the remaining hospitals in the local health care market (Lindrooth, Lo Sasso, & Bazzoli, 2003). Understanding causes of hospital closure is important if hospitals are to survive and continue to fulfill their missions as the center for health care among their neighborhoods. For hospital administrators and others interested in hospital performance, knowing which hospitals are most susceptible to closure can be of great use.
Researchers have identified many factors associated with increased risk of hospital closure. Internal structural factors of a hospital such as bed size (Lee & Alexander, 1999; Lillie-Blanton et al., 1992; Lindrooth et al., 2003; Longo, Sohn, & Shortell, 1996; Succi, Lee, & Alexander, 1997), ownership type (Succi et al., 1997; Williams, Hadley, & Pettengill, 1992), and teaching status (Lee & Alexander, 1999) were predictive of hospital closure. Increased risk of hospital closure was found among hospitals located in certain areas where area socioeconomic status indicators were not good (Lee & Alexander, 1999; Longo & Chase, 1984) and/or inter-hospital competition was harsh (Lee & Alexander, 1999; Longo & Chase, 1984; Longo et al., 1996; Mayer, Kohlenberg, Sieferman, & Rosenblatt, 1987; Mullner, Rydman, Whiteis, & Rich, 1989; Succi et al., 1997). Process variables such as occupancy rate were suggested as a mediator between these risk factors and the outcome of hospital closure (Kennedy & Dumas, 1983; Lindrooth et al., 2003; Longo & Chase, 1984; Lynch & Ozcan, 1994). In addition, some researchers have explored the preventive effect of hospitals’ strategic actions on hospital closure (Lee & Alexander, 1999; Longo et al., 1996; Succi et al, 1997). However, most of these studies are American and do not directly relate to the health care systems in other countries.
With the establishment of national health insurance in 1989, the demand for hospital services in South Korea has significantly risen with the increasing number of hospitals. Although the health insurance has supported South Korean hospitals financially, the South Korean medical community has claimed that the level of fee schedule is low. In addition, the high rate of hospital closure has been a major source of concern for the hospital sector. According to a report, 8.1% of all hospitals in 2001 were in the status of bankruptcy (Yang, 2002). Several Korean studies were performed to examine risk factors of hospital closure and to predict the hospital closure (Jung & Lee, 2000; Lee & Seo, 1998; Yang, 2002). Several financial indicators (i.e., profitability, liquidity, activity measures) were predictive of the hospital closure. However, these studies have limitations regarding the representativeness of the sample studied and failed to consider the environmental risk factors like inter-hospital competition.
Viability of a hospital is less likely determined by hospital distress at one time point. Closure of a hospital would be the result of successive long-term exposure of organizational and/or environmental risk factors rather than from cross-sectional exposure of those factors during short time periods. Several studies included information on exposure of risk factors at multiple time points and employed generalized estimating equations or Cox's regression model (Alexander, D’Aunno, & Succi, 1996; Lee & Alexander, 1999; Succi et al., 1997). In studies using longitudinal data with repeated measurement with the same subjects, correlation among these repeated observations should be considered in the analysis. In order to control correlation among repeated measurements, in this study we used a frailty model. The frailty model is an extension of Cox's model which allows frailties or random effects. The hazard function for each hospital may depend on observed risk factors, but usually not all such factors are known or measurable. This unknown factor of the hazard function is usually termed the frailty or random effect. When the repeated measurements of a particular type of event (closure) are obtained from the same hospital, frailty is an unobserved common factor for each hospital and is thus responsible for creating the dependence between repeated measures. This frailty is often regarded as a random quantity from some suitably defined population distribution of frailties. We used hierarchical generalized linear model (HGLM, Lee & Nelder (1996), Lee & Nelder (2001)) to implement the frailty model and examined determinants of hospital closure in a nationally representative data sample in South Korea. We hypothesized that environmental factors like hospital competition as well as hospital characteristics like ownership play an important role in determining hospital closure in South Korea.
Section snippets
Definition of hospital closure and market area
A hospital experiences many levels of organizational change between survival and closure (Alexander et al., 1996). In this study, the status of all hospitals was categorized as closure, survival, and censored. According to South Korea's Medical Law, hospitals (including general hospitals) need to have a minimum of 30 inpatient beds (Organization for Economic Co-operation and Development (OECD), 2002). Medical institutions that do not meet this requirement remain classified as clinics. We
Results
As shown in Table 2, 203 (25.2%) of 805 hospitals closed between 1996 and 2002. The annual closure rate in 1998 was 5.1% but decreased to 2.8–3.7% in 2001–2002. The average annual closure rate was 4.2%. The censored case (functional changes to clinics) during the study period was 26 (3.2%).
Table 3 presents baseline characteristics of study hospitals and their neighborhoods in 1996. Most hospitals were private (91.1%) and located in urban areas (87.8%). The mean number of beds was 185 and the
Discussion
Between 1996 and 2002, the average annual hospital closure rate in South Korea was 4.2%. This is about 3 times higher than that of the US (1.3%) in 1999 (Office of Inspector General, 2001). This high rate of hospital closure has been a major source of concern for the South Korean medical community. While many studies have investigated determinants for individual hospital closures (Jung & Lee, 2000; Lee & Seo, 1998; Yang, 2002), this study is the first to examine determinants of hospital closure
References (38)
- et al.
The closure of rural hospitals in Saskatchewan: Method or madness?
Social Science & Medicine
(2001) - et al.
The effect of urban hospital closure on markerts
Journal of Health Economics
(2003) - et al.
Patterns of rural hospital closure in the United States
Social Science & Medicine
(1987) - et al.
Determinants of profound organizational change: Choice of conversion or closure among rural hospitals
Journal of Health and Social Behavior
(1996) - et al.
A public hospital closes: Impacts on patients’ access to care and health status
Journal of the American Medical Association
(1990) - et al.
Measuring geographic access to health care in rural areas
Medical Care Review
(1994) Defining rural hospital markets
Health Services Research
(1993)- et al.
Effects of hospital closure on mortality rates of the over-65 long-stay psychiatric population
International Journal of Geriatric Psychiatry
(1998) - et al.
Developing a combined forecasting model on hospital closure
Korean Journal of Health Policy and Administration
(2000) - et al.
Marginal likelihoods based on Cox's regression and life model
Biometrika
(1973)
Hospital closures and survivals: An analysis of operating characteristics and regulatory mechanisms in three states
Health Services Research
Managing hospitals in turbulent times: Do organizational changes improve hospital survival?
Health Services Research
Hierarchical generalized linear model
Journal of the Royal Statistical Society B
Hierarchical generalized linear models: A synthesis of generalized linear models, random-effect models and structured dispersions
Biometrika
Predicting hospital bankruptcy in Korea
Korean Journal of Preventive Medicine
Rural and urban hospital closures, 1985–1988: Operating and environmental characteristics that affect risk
Inquiry
Structural determinants of hospital closure
Medical Care
Cited by (12)
The impact of hospital closures on geographical access: Evidence from four southeastern states of the United States
2017, Operations Research PerspectivesCitation Excerpt :One of the possibilities includes the closure of hospitals. Whether or not hospitals are private or public, ownership (and hospital size) appear to makes a difference when considering the closure of hospitals, as noted by Noh et al. [9] in the case of Korea. Furthermore, consolidation among private health care companies often leads to closure of hospitals [10].
Rural vs urban hospital performance in a 'competitive' public health service
2010, Social Science and MedicineCitation Excerpt :Braithwaite (1997) argues that a market runs the risk of failing to provide healthcare services equitably in publicly funded systems. The introduction of market-driven reforms has resulted in the closure of hospitals, especially in rural areas, in the US (Adams, Porell, & Robbins, 1996; GAO, 1990) Korea (see Noh et al., 2006), Canada (see James, 1999; Lepnurm & Lepnurm, 2001) and New Zealand (see Healy, 2002). However, in Western Europe, the restructuring of acute care has meant a reduction in the number of acute beds in hospitals (Kroneman & Siegers, 2004) but only rarely has it led to the closure of hospitals (Healy & McKee, 2002).
Seeking affective health care: Korean immigrants' use of homeland medical services
2010, Health and PlaceAssessing the Financial Health of Hospitals Using a Financial Index
2018, Springer Proceedings in Business and EconomicsIncreasing the productivity of health facilities in urban areas in México
2018, Proceedings of the International Conference on Industrial Engineering and Operations ManagementDefining the activities of publicness for Korea's public community hospitals using the Delphi method
2017, Medicine (United States)
- ☆
This study was supported by a grant of the Korea Health 21 R&D Project, Ministry of Health & Welfare, Republic of Korea (01-PJ1-PG3-51200-0002).