Dynamics of ambulatory surgery centers and hospitals market entry

Abstract

In this article, we investigate the diversity of healthcare delivery organizations by comparing the market determinants of hospitals entry rates with those of ambulatory surgery centers (ASCs). Unlike hospitals, ASCs is one of the growing populations of specialized healthcare delivery organizations. There are reasons to believe that firm entry patterns differ within growing organizational populations since these markets are characterized by different levels of organizational legitimacy, technological uncertainty, and information asymmetry. We compare the entry patterns of firms in a mature population of hospitals to those of firms within a growing population of ASCs. By using patient-level datasets from the state of Florida, we break down our explanatory variables by facility type (ASC vs. hospital) and utilize negative binomial regression models to evaluate the impact of niche density on ASC and hospital entry. Our results indicate that ASCs entry rates is higher in markets with overlapping ASCs while hospitals entry rates are less in markets with overlapping hospitals and ASCs. These results are consistent with the notion that firms in growing populations tend to seek out crowded markets as they compete to occupy the most desirable market segments while firms in mature populations such as general hospitals avoid direct competition.

Keywords

competition hospitals market entry niche overlap surgery centers

Introduction

Countries in Europe and other parts of the world have to a various extent shifted towards a regulated competition model in both the healthcare providers and health insurance markets.¹ While internationally this move towards competition and contestability is a relatively recent trend, the United States has the longest and broadest experience in competitive markets. As a byproduct of its relatively competitive markets, starting from the early 1990s, the United States has witnessed the market entry and proliferation of new organizational forms such as urgent care centers, ambulatory surgery centers (ASCs), and specialty hospitals. This increasing diversity in healthcare delivery organizations in the United States can be best described as the growth of specialized facilities and the contraction of the traditional “do-it all” general hospitals.^2–5 This trend is not restricted to the United States, other countries has also witnessed an increase in the number of specialized healthcare delivery organizations. In the United Kingdom, specialty hospitals have been encouraged by the National Health Service (NHS) for the potential gains they offer in terms of efficiency and quality.⁶

In the United States, in addition to the proliferation of specialty hospitals,³ ASCs’ growth might be the fastest among new specialized healthcare delivery organizations. In 2005, there were 4445 ASCs in the United States,⁷ while there were only 400 in 1983 and 3300 in 2001.⁸ ASCs capitalized on the shift away from inpatient towards outpatient procedures. They perform surgical procedures for which an overnight hospital stay is not required and could be multi-specialty centers or single specialty, with ophthalmologic, gastroenterologic, and orthopedic centers being the most dominant types.⁹

Competition and market entry of healthcare delivery organizations are important for policymakers as well as managers worldwide. As governments continuously seek to shape their healthcare markets in ways that encourage efficiency and improve quality, it is imminent to understand the driving forces that impact market entry rates of new organizational forms, specifically specialized healthcare delivery organizations or “focused factories.” Research from other industries indicates that mature organizational populations such as general hospitals behave differently from growing organizational populations such as specialty hospital and ASCs. Moreover, theory from other sectors suggests that entry decisions of growing populations are influenced by different if not opposing forces from market entry decisions of mature populations such as hospitals.¹⁰ One of these key forces that are worth studying given its policy implications is competition. The majority of research in healthcare treats competition as attribute of a defined market. However, organizations within the same market are faced by different levels of competition based on the level of overlap in the services they provide.¹¹ In this paper, we rely on the niche overlap theory perspective to investigate the impact of relational competition on the market entry of hospitals and ASCs. We also provide a framework to differentiate between market entry into growing or emerging organizational populations and mature populations in healthcare delivery markets. Specifically, we rely on r- and K-selection processes to investigate the market entry into the growing population of ASCs and the mature population of hospitals. We use patient-level datasets from the state of Florida to evaluate the impact of niche density on ASCs and hospital entry rates.

Theoretical framework

Research on the determinants of market entry of healthcare organizations is not abundant. In general theories explaining market entry rates in healthcare can be grouped under two general categories: (1) health economics and (2) organizational ecology.¹² The health economics perspective focuses on forces of supply and demand and market competition.¹³ However, the organization ecology perspective focuses on populations of organizations that occupy different niches and the factors that influence the market entry and exit of different organizational populations.¹² In this paper, we incorporate both the health economics perspective by controlling for forces of supply and demand and the organizational ecology by examining two organizational populations: ASCs and general hospitals. Our theoretical framework relies on niche overlap theory and extends the literature on market entry by incorporating the r- and K-selection processes. The main contribution of our model is that it provides a framework to study market entry by integrating competition between organizations with the same form and competition between organizations with different forms, i.e., generalist and specialists. Moreover, our model provides a theoretical framework that distinguishes between entry into mature markets such as hospitals versus entry into emerging markets such ASCs. Our model provides a comprehensive framework for studying the emergence and rise of new organizational forms and the interaction between emerging organizational forms and incumbent ones. Finally, we do not limit our model to competition but we also incorporate other dynamics such as mutualism.

Niche overlap theory

Traditionally, competition in healthcare markets is treated as an attribute of a defined market whereby all hospitals in the market face the same level of competition.¹¹ Niche overlap theory, however, provides a relational view of competition where competition is measured by the level of overlap in the resource niches between two hospitals which allows for a more fine-grained conceptualization of competition faced by hospitals.¹¹ According to the niche overlap theory, every organization in a population occupies an organizational niche characterized by a location in resource space.^14,15 Depending on the organizational niches they target, organizations face different competitive landscapes. Organizations that operate in the same organizational niche experience competitive effects since they compete directly for scarce resources. Organizations that occupy non-overlapping organizational niches experience mutualistic effects by, for example, cooperating directly or providing services that create complementary demand. Therefore, the level of competition between two ASCs or two hospitals is proportional to the overlap in the required resources rather than a concentration measure that characterizes all hospitals in their market.¹⁶

Baum and Singh¹⁵ argued that the degree of organizational niche overlap and non-overlap also influences the organizational niches in which entrepreneurs choose to create organizations. They define overlap density as the total number of firms with overlapping niches, i.e., providing the same services or products. Higher overlap density signals to entrepreneurs high level of competition. Organizations targeting niches that are more crowded have a much lower likelihood of being founded than organizations targeting less densely niches. More precisely, the higher the overlap density, the greater is the intensity of competition, and the lower the entry rate. Therefore, consistent with classical economic theory, niche overlap theory suggests that hospitals are less likely to enter into markets where there is a high level of overlap in the services already provided.

Hypothesis 1: Overlap density is negatively related to the entry rate of hospitals.

Likewise, they defined niche non-overlap density to be the total number of firms with non-overlapping niches.¹⁵ Non-overlap density captures the number of organizations in the population that do not offer the same services or have the same resource requirement as the focal organization. Baum and Singh argued that non-overlap density has the opposite effect on market entry. Non-overlap density captures information about differentiation of organizational niches. This differentiation leads to complementarities between organizations that occupy different niches and thus create mutualistic interdependencies among them. For example, organizations benefit from the broader social acceptance and greater institutional support posed by non-overlapping firms with the same organizational population. Likewise, there exist opportunities for demand enhancement and other cooperative endeavors among non-overlapping organizations. The confluence of these factors suggests that non-overlap density will have a positive effect on the firm entry rate.

By examining both mutualistic effects and relational competition, niche overlap theory provides a different perspective in treating competition and market entry in the healthcare delivery markets. The niche overlap theory has found support regarding the impact of niche overlap and non-overlap on firm entry and exit rates within a variety of settings including daycare centers and the automobile industry.^15,17–19 Yet, while niche overlap found support in the literature, the theory remains limited in its scope since it emphasize micro-niches and had not yet explored the effect of niche overlap and non-overlap between organizational populations occupying different macro-niches—specifically, specialist and generalist organizational forms. Therefore, we draw from existing theory to pose some hypotheses regarding the effects of competition between these two organizational forms at different stages of industry evolution. Specifically, we rely on the r- and K-selection processes and density dependence theories.

R and K-selection and entry strategies

Previous research on market entry of healthcare organizations has ignored differences between more established forms such as hospitals and new organizational forms such as ASCs. Brittain and Freeman¹⁰ point out that organizations frequently pursue different entry strategies, depending on the stage of industry evolution. In order to justify their theory, they draw from models used by biologists to model the growth of populations in finite environments:

\frac{d X}{dt} = rX (\frac{K - X}{K})

(1.1)

where the rate of change of a population with size X is a function of some rate of nature increase r, and carrying capacity K. Ecologists sometimes describe alternative modes of selection—or strategies for survival—in terms of the parameters of this equation: r and K.²⁰

Pure r-strategists are organizations that move quickly to exploit resources as they first become available.¹⁰ Their structure makes them relatively inexpensive to set up since they concentrate on activities that require low levels of capital investment and simple structures. Their success depends heavily on first-mover advantages, which makes them high-risk and high payoff organizations that gain maximally from temporarily rich environments. K-strategists, on the other hand, are organizations that are structured to compete successfully in densely settled environments.¹⁰ K-strategist organizations generally expand more slowly into new resource spaces than r-strategists because the structures generating competitive efficiency frequently preclude the rapid adjustments necessary to capture first-mover advantages. Competition on the basis of efficiency generally involves higher levels of investment in plant and equipment and more elaborate organizational structures.

In industries with changing products, developing technologies, or rapidly increasing market munificence, competition is dominated by r-selection which favors emerging organizations capable of moving quickly to exploit new resource opportunities in the resource rich but uncertain environments that characterize low-density conditions. As the emerging population continues to grow, its markets become connected and environmental uncertainty is reduced as demand becomes more predictable. As an industry stabilizes and the organizational population reaches its carrying capacity, the resources available to its members are exploited fully and competitive pressures shift to K-selection, which favors organizations competing on the basis of efficiency.

Therefore, the r- and K-selection processes may produce different patterns of firm entry into organizational niches. In the case of mature populations, such as general hospitals, characterized by K-selection processes, organizations compete on the basis of efficiency¹⁰ and these competitive pressures increase within markets that are more crowded. Thus, general hospital should avoid direct competition with other firms and favor uncrowded niches. Yet this differs within a growing population of a new organizational form such as ASCs. The theory suggests that r-selectionists like ASCs will compete on the basis of market positioning and will seek out the most profitable markets regardless of whether they are relatively crowded or not. Thus, surgery centers will enter markets in which there is high niche overlap density. Surgery centers represent an excellent example of a growing organizational population since the rapid increase in the number of facilities indicates that it is characterized by r-selection processes. Meanwhile, the hospital sector has existed for centuries and the fact that the number of U.S. hospitals is declining would suggest that the organizational population is relatively mature and characterized by K-selection processes.

Hypothesis 2: Overlap density is positively related to the entry rate of ASCs.

Methods

We chose the state of Florida as our study site for several reasons. First, it is the state with the second highest number of ASCs in the United States. In fact, the number of surgery centers in Florida grew from just 110 in 1993 to 379 in 2006, so it closely mirrors the trends in ASC growth being observed throughout the country. More importantly, the state of Florida requires that healthcare facilities report data on all inpatient and outpatient surgical procedures that they perform. Patient data reported by Florida healthcare facilities is compiled by the Agency for healthcare administration (AHCA) and represents a complete census of all inpatient and outpatient surgeries that have occurred within the state from 1997 to 2006. For the inpatient data, surgical procedures are represented by ICD-9-CM/ICD-10-CM procedure codes. Within the outpatient data, surgical procedures are represented by CPT/HCPCS procedure codes. In order to transform these patient datasets into quarterly procedure counts, we coded these procedures and categorized them within surgical specialties. Clinical classification software (CCS) from the Healthcare utilization project (HCUP) allowed us to classify both ICD-9-CM/ICD-10-CM and CPT/HCPCS codes among 244 different procedure types. Cross-tabulating these procedures types with facility IDs and patient county codes produced two panel datasets in which the unit of observation is the facility-quarter and county-quarter while each column represents that facility and county’s procedure count within each of the 15 surgical specialties, respectively. We modeled facility entry as a function of each specialty’s characteristics within each county. We subsequently converted these datasets from wide to long datasets so that the each unit of observation was the facility-specialty-quarter and county-specialty-quarter. Our dependent and independent variables were calculated from these two datasets.

Conveniently, the state of Florida maintains data on the location of all surgery centers and general hospitals, which allows us to assign them to their county of residence. Each of these 67 counties was subsequently assigned to one of 17 health service areas (HSAs) based upon the market boundaries established by the National Center for Health Statistics.²¹ We defined these HSAs to be their own markets for the purposes of this study, and all explanatory variables were initially measured at the HSA level. Although HSAs provide a convenient starting point from which to define markets, they span very large geographic areas. Therefore, we elaborated upon them in order to distinguish between local and diffuse market conditions since travel distances clearly play a role in the market for outpatient services.^21,22 We calculated our local measures from the county within which each surgery center resides and our diffuse measures from all other counties within that surgery center’s HSA. Our explanatory variables were also divided and calculated separately in the same manner.

Dependent variables

We began our analysis with the dataset of quarterly procedure counts that we had tabulated by facility ID codes in which the unit of observation is the facility-quarter. Facilities enter this panel dataset as they enter the market for outpatient services and so this panel provided us a starting point from which to observe facility entry. The state of Florida also maintains information on all licensed and accredited outpatient facilities in the state. This dataset extends back to the early-1990s and provides the date(s) of opening for all such entities. This licensure dataset was merged with our outpatient procedure dataset through AHCA ID numbers in order to validate the patterns of activity produced by our panel dataset with the state’s official entry. From these datasets, we identified a total of 193 entries among 406 ASCs and 10 entries among 222 hospitals from 1997 to 2006, which represent fairly reasonable sample sizes when multiplied by the number of specialties that each of these facilities enter over 40 quarters.

We assessed the causes of facility entry not simply into counties but rather into each specialty within those counties as a function of those county-specialty characteristics. The county-specialty is the unit of observation for the dependent variable in our empirical models while our independent variables were initially measured at the HSA level and were subsequently broken down by facility type (ASC vs. hospital) and geographic location (local vs. diffuse). In order to assess whether dependent and independent variables should be calculated from the same geographic area, we ran separate analyses with all variables measured at the county and HSA level to check the robustness of our results.

Independent variables

In order to calculate competition at the county-specialty level, we divided each facility up by the proportion of its total procedure volume that it devoted to each surgical specialty. Following the example of Baum and Singh,¹⁵ we constructed niche overlap weights from the proportion of each facility’s procedure volume that was devoted to specialty j:

w_{ij} = \frac{s_{ij}}{\sum_{j = 1}^{15} s_{ij}} w

(1.2)

where s_j represents the number of procedures that facility i performs within specialty j. Given these weights, the overlap and non-overlap density within each county-specialty was calculated from the following formulas:

overlap {density}_{jk} = \sum_{i = 1}^{n} w_{ij} non - overlap {density}_{jk} = n - \sum_{i = 1}^{n} w_{ij}

(1.3)

where w_ij represents the weights that were described previously and n represents the number of outpatient facilities within county k. Adding up niche overlap density across all surgical specialties yielded the total number of outpatient facilities within each county. By subtracting niche overlap density from the total number of outpatient facilities in each county, we were able to calculate niche non-overlap density. We calculated these measures separately by facility type (ASC vs. hospital) and geographic location (local vs. diffuse) in order to evaluate the differential impact of each. We also included variables representing the size of demand within each county-specialty, which represents the environmental carrying capacity. Quarterly procedure counts were deemed to be a reasonable proxy for the level of demand within each market. We generated these quarterly procedure counts from both the inpatient and outpatient datasets, which produced two separate sets of procedure counts; we included outpatient procedure counts within our entry models for ASCs and hospitals while we included the inpatient counts for hospitals only. As with our measures of competition, each of these demand statistics was broken down by facility type (ASC vs. hospital) and geographic location (local vs. diffuse).

We also included a variety of control variables at the county and county-specialty level. For county-specialty controls, we included payer mix participation in order to account for the possibility that certain specialties are reimbursed more generously and therefore are more appealing. Although the Florida outpatient dataset utilizes 16 different payer codes, we collapsed these codes into three categories: public, private, and other. We included variables representing the proportion of a facility’s business that is reimbursed by public and private payers while omitting the variable representing other payers as a comparison group. We broke these variables down in order to observe whether they were dependent upon geographic location (local vs. diffuse).

Relying on data from Florida’s AHCA, we calculated the proportion of Medicaid enrollees within each HSA and then broke these figures down by geographic location (local vs. diffuse). We also included a number of annual demographic and economic variables in order to model facility entry as a function of general environmental munificence. We obtained data on the proportion of minority residents and population size within each county from the U.S. Census Bureau and divided up these variables by age (0–64 and 65+). We also included variables representing the per capita income for the residents within each county and these estimates were obtained from the U.S. Bureau of Economic Analysis Local Area Personal Income files. Finally, we included estimates of the proportion each county’s population living below the federal poverty line from the U.S. Census Bureau’s Small Area Income and Poverty Estimates. Table 1 presents the summary statistics for the variables included in our models.

Table 1.

Summary statistics.

Category	Facility type	Location	Variable	Model	Mean	SD	Min	Max
Competition	Overlap density	1	1.799	3.302	0.000	27.663
	Non-overlap density	1	25.184	20.123	1.548	89.766
	ASC	Overlap density	2	1.012	2.190	0.000	18.290
	ASC	Non-overlap density	2	14.165	11.624	0.000	59.000
	Hospital	Overlap density	2	0.787	1.464	0.000	16.934
	Hospital	Non-overlap density	2	11.019	9.672	0.407	46.904
	Local	Overlap density	3	0.446	1.164	0.000	16.155
	Local	Non-overlap density	3	6.244	8.467	0.000	44.894
	Diffuse	Overlap density	3	1.353	2.599	0.000	26.211
	Diffuse	Non-overlap density	3	18.940	16.388	0.000	82.782
	ASC	Local	Overlap density	4	0.251	0.792	0.000	10.009
	ASC	Local	Non-overlap density	4	3.520	4.950	0.000	24.997
	ASC	Diffuse	Overlap density	4	0.760	1.747	0.000	16.636
	ASC	Diffuse	Non-overlap density	4	10.646	9.752	0.000	55.000
	Hospital	Local	Overlap density	4	0.195	0.514	0.000	9.146
	Hospital	Local	Non-overlap density	4	2.725	3.992	0.000	25.953
	Hospital	Diffuse	Overlap density	4	0.592	1.125	0.000	15.520
	Hospital	Diffuse	Non-overlap density	4	8.294	7.489	0.000	43.904
Outpatient demand	Number of procedures	1	2225.093	4213.847	1.000	47304.000
	ASC	Number of procedures	2	956.348	2252.075	0.000	24954.000
	HOSP	Number of procedures	2	1268.745	2389.245	1.000	26110.000
	Local	Number of procedures	3	565.234	1525.518	0.000	26345.000
	Diffuse	Number of procedures	3	1659.859	3280.609	0.000	46448.000
	ASC	Local	Number of procedures	4	242.719	779.355	0.000	14934.000
	ASC	Diffuse	Number of procedures	4	713.629	1762.360	0.000	24551.000
	Hospital	Local	Number of procedures	4	322.515	903.378	0.000	15672.000
	Hospital	Diffuse	Number of procedures	4	946.230	1850.197	0.000	25320.000
Inpatient demand	Number of procedures	1,2	1102.304	1989.015	0.000	15135.000
	Local	Number of procedures	3,4	278.277	726.329	0.000	8906.000
	Diffuse	Number of procedures	3,4	824.027	1546.030	0.000	14941.000
Payer type	Percent public pay	1	0.438	0.163	0.000	0.899
	ASC	Percent public pay	2	0.350	0.236	0.000	1.000
	Hospital	Percent public pay	2	0.452	0.154	0.000	0.880
	Local	Percent public pay	3	0.460	0.220	0.000	1.000
	Diffuse	Percent public pay	3	0.431	0.183	0.000	1.000
	ASC	Local	Percent public pay	4	0.324	0.294	0.000	1.000
	ASC	Diffuse	Percent public pay	4	0.340	0.249	0.000	1.000
	Hospital	Local	Percent public pay	4	0.468	0.226	0.000	1.000
	Hospital	Diffuse	Percent public pay	4	0.445	0.177	0.000	1.000
	Percent private pay	1	0.508	0.157	0.064	1.000
	ASC	Percent private pay	2	0.529	0.264	0.000	1.000
	Hospital	Percent private pay	2	0.498	0.149	0.083	1.000
	Local	Percent private pay	3	0.466	0.214	0.000	1.000
	Diffuse	Percent private pay	3	0.487	0.182	0.000	1.000
	ASC	Local	Percent private pay	4	0.422	0.325	0.000	1.000
	ASC	Diffuse	Percent private pay	4	0.499	0.284	0.000	1.000
	Hospital	Local	Percent private pay	4	0.457	0.220	0.000	1.000
	Hospital	Diffuse	Percent private pay	4	0.477	0.175	0.000	1.000
	Percent other pay	1	0.054	0.063	0.000	0.540
	ASC	Percent other pay	2	0.060	0.106	0.000	1.000
	Hospital	Percent other pay	2	0.050	0.058	0.000	0.471
	Local	Percent other pay	3	0.050	0.081	0.000	1.000
	Diffuse	Percent other pay	3	0.051	0.064	0.000	1.000
	ASC	Local	Percent other pay	4	0.055	0.124	0.000	1.000
	ASC	Diffuse	Percent other pay	4	0.056	0.106	0.000	1.000
	Hospital	Local	Percent other pay	4	0.046	0.082	0.000	1.000
	Hospital	Diffuse	Percent other pay	4	0.048	0.059	0.000	1.000
Market controls	Population (0–64)	1,2	813800.000	795476.700	121779.000	3660695.000
	Local	Population (0–64)	3,4	202184.900	340859.600	5550.000	2060359.000
	Diffuse	Population (0–64)	3,4	611615.100	625181.800	0.000	3597333.000
	Population (65+)	1,2	163395.100	155885.900	19129.000	603917.000
	Local	Population (65+)	3,4	42600.350	64447.820	675.000	341849.000
	Diffuse	Population (65+)	3,4	120794.800	123492.600	0.000	592505.000
	Percent minority	1,2	0.276	0.104	0.100	0.677
	Local	Percent minority	3,4	0.254	0.137	0.064	0.815
	Diffuse	Percent minority	3,4	0.264	0.118	0.000	0.798
	Per capita income	1,2	27764.660	6393.681	17301.900	50549.290
	Local	Per capita income	3,4	24887.300	8238.906	12032.590	57445.930
	Diffuse	Per capita income	3,4	26679.030	8298.582	0.000	54535.170

Empirical model

Our empirical models estimate the probability of facility entry as a function of market characteristics which include overlap density, non-overlap density, demand size, and market control variables. We modeled our dependent variable, the quarterly number of entries within a county-specialty in quarter t, as a function of our independent variables in quarter t–1.

We utilized negative binomial regression to estimate our model but we also validated the robustness of our results by running a Poisson regression model as well. One of the assumptions of the negative binomial model is that all the observations be independent. However, we know that facility entries within a county-specialty may not be independent since multi-specialty facilities enter more than one specialty at a given point in time. So we clustered our standard errors by county in order to allow for intragroup correlation, which relaxed the requirement that the observations be independent. We also included year and quarter dummy variables in order to control for any time trends present in the data.

Since we broke down each category of explanatory variables by facility type (surgery center vs. hospital) and geographic location (local vs. diffuse), we introduced these distinctions in a relatively gradual manner. The first model included every set of variables at the highest possible level, the second model broke them down by facility type (ASC vs. hospital), the third model broke them down by geographic location (local vs. diffuse), and the fourth model broke them down both by facility type and geographic location. We estimated our negative binomial models of facility entry with STATA version 10 and ran the models separately for ASCs and hospitals. Tables 2 and 3 present the results of our negative binomial models; variables from different models have been lined up horizontally rather than vertically so that similar elements are grouped together.

Table 2.

Surgery center entry—negative binomial model.

Category	Facility type	Location	Measure	Model 1	Model 2	Model 3	Model 4
Category	Facility type	Location	Measure	Model 1	Facility	Geography	Facility Geography
Competition	ASC	Local	Overlap	0.03	0.0267	0.00854	0.0597
		Local	Non-overlap	−0.0434	−0.0456	−0.0174	−0.0218
		Diffuse	Overlap	0.174***	0.144***
		Diffuse	Non-overlap	0.0239	0.0359
	Hospital	Local	Overlap	−0.00126	−0.185
		Local	Non-overlap	−0.137*	−0.126*
		Diffuse	Overlap	0.166*
		Diffuse	Non-overlap				−0.0868
Procedure demand	ASC	Local	Number of procedures	−5.37E–06	−1.87E–05	4.27E–05	−1.45E–05
	ASC	Diffuse	Number of procedures	−7.87e–05**	−6.88E–05
	Hospital	Local	Number of procedures	−3.81E–05	7.21E–05
	Hospital	Diffuse	Number of procedures	−0.000122**
Payer type	ASC	Local	Percent public pay	−2.322***	1.280***	−0.748*	0.852***
	ASC	Diffuse	Percent public pay	0.0827	0.958***
	Hospital	Local	Percent public pay	−4.502***	−1.185***
	Hospital	Diffuse	Percent public pay	−0.389
	ASC	Local	Percent private pay	−2.989***	0.912***	−0.43	0.931***
	ASC	Diffuse	Percent private pay	−0.853	0.604**
	Hospital	Local	Percent private pay	−4.861***	−0.793**
	Hospital	Diffuse	Percent private pay				−0.962
Market controls	Local	Population (0–64)	1.62e–06*	2.00e–06**	3.13e–06***	4.14e–06***
	Diffuse	−2.25E–07	2.60E–07
	Local	Population (65+)	2.38E–06	7.49E–06	−9.91E–07	2.73E–06
	Diffuse	−3.48E–06	9.18E–07
	Local	Percent minority	−6.418**	−6.755**	−4.019***	−4.626***
	Diffuse	−1.435	−1.187
	Local	Per capita income	8.74e–05**	6.39E–05	6.28e–05***	4.80e–05***
	Diffuse	1.97E–05	5.26E–06
Model controls	Quarter 2 dummy	−0.232	−0.221	−0.231	−0.222
	Quarter 3 dummy	−0.146	−0.128	−0.143	−0.137
	Quarter 4 dummy	−0.297	−0.29	−0.302	−0.305
	Year 1998 dummy	0.0222	0.0562	−0.027	−0.0373
	Year 1999 dummy	−0.319	−0.456	−0.404	−0.601
	Year 2000 dummy	−0.871	−1.038*	−0.933	−1.198**
	Year 2001 dummy	−0.980*	−1.103**	−1.075**	−1.294**
	Year 2002 dummy	−0.242	−0.363	−0.371	−0.601
	Year 2003 dummy	−0.232	−0.337	−0.416	−0.645
	Year 2004 dummy	−0.269	−0.363	−0.524	−0.742
	Year 2005 dummy	−1.045	−1.092	−1.358**	−1.542***
	Year 2006 dummy	−1.139	−1.187	−1.513**	−1.694**
			Constant	−2.282*	−0.543	−3.873***	−3.646***

Note: Entry dates validated by licensure data. Standard errors clustered by county. Time-varying independent variables lagged by one quarter.

***

p < 0.01, **p < 0.05, *p < 0.1.

Table 3.

Hospital entry—negative binomial model.

Category	Facility type	Location	Measure	Model 1	Model 2	Model 3	Model 4
Category	Facility type	Location	Measure	Model 1	Facility	Geography	Facility Geography
Competition	ASC	Local	Overlap	−0.230*	−0.400***	−0.902***	−0.712**
		Local	Non-overlap	−0.088	−0.0923	−0.421***	−0.156
		Diffuse	Overlap	0.104	−0.163
		Diffuse	Non-overlap	0.202	0.138
	Hospital	Local	Overlap	−0.533	−2.056**
		Local	Non-overlap	−0.616**	−2.493**
		Diffuse	Overlap	0.384
		Diffuse	Non-overlap	0.292
Outpatient demand	ASC	Local	Number of procedures	0.000114	0.000270*	0.000297**	0.000423
	ASC	Diffuse	Number of procedures	0.000105	0.000277
	Hospital	Local	Number of procedures	−5.38E–05	−0.000396
	Hospital	Diffuse	Number of procedures				6.85E–06
Inpatient demand	Hospital	Local	Number of procedures	−6.07E–06	−9.45E–06	2.26E–05	0.000121*
Inpatient demand	Hospital	Diffuse	Number of procedures	−4.49E–05	−3.15E–05
Payer type	ASC	Local	Percent public pay	−2.096	0.516	−0.432	0.162
	ASC	Diffuse	Percent public pay	−2.712	1.642*
	Hospital	Local	Percent public pay	−2.628	−0.422
	Hospital	Diffuse	Percent public pay	−5.182***
	ASC	Local	Percent private pay	−3.183	−0.805*	−0.633	−0.615**
	ASC	Diffuse	Percent private pay	−5.018**	1.009*
	Hospital	Local	Percent private pay	−2.663	0.103
	Hospital	Diffuse	Percent private pay				−7.166***
Market controls	Local	Population (0–64)	3.16e–06*	6.10e–06***	9.72e–06***	2.09e–05***
	Diffuse	1.94E–06	2.45E–06
	Local	Population (65+)	−1.62E–06	1.97E–05	1.95E–05	6.94e–05**
	Diffuse	−3.99e–05**	−4.12e–05**
	Local	Percent minority	−7.405	−6.008	−5.459	−2.635
	Diffuse	−6.816	−6.903
	Local	Per capita income	5.27E–05	−4.15E–05	8.38E–05	−1.79E–05
	Diffuse			6.86E–05	0.000157***
Model controls	Quarter 2 dummy	−0.581	−0.515	−0.644	−0.643
	Quarter 3 dummy	−0.319	−0.227	−0.466	−0.417
	Quarter 4 dummy	−0.989	−0.888	−1.151	−1.088
	Year 1998 dummy	1.705	1.109	2.805	2.177
	Year 1999 dummy	−14.64***	−14.73***	−12.65***	−13.68***
	Year 2000 dummy	−14.75***	−14.76***	−12.77***	−13.65***
	Year 2001 dummy	1.468	1.504	2.592	2.668
	Year 2002 dummy	2.394*	2.291*	3.371**	3.350*
	Year 2003 dummy	1.53	1.465	2.366	2.43
	Year 2004 dummy	1.471	1.433	2.039	1.899
	Year 2005 dummy	2.259*	2.314*	2.633**	2.876**
			Constant	−3.867	−1.258	−5.817	−6.095

Note: Entry dates validated by licensure data. Standard errors clustered by county. Model excludes observations from 1997. Time-varying independent variables lagged by one quarter.

***

p < 0.01, **p < 0.05, *p < 0.1.

Results

Our negative binomial models yielded log likelihood statistics that were highly significant (p < 0.0001). Our models generally produced coefficients with the expected signs, but we ran Poisson models to check the robustness of our results and found that none were substantively different. Higher levels of market demand for outpatient procedures are associated with a higher probability of ASC entry (Models 1 to 4, p < 0.05), while hospital entry is only positively associated with outpatient demand from ASCs (Model 4, p < 0.05) and inpatient demand (Model 2, p < 0.05).

In terms of market control variables, ASC entry rates were significantly higher when ASCs are reimbursed predominantly by public and private payers (Models 2 and 4, p < 0.05) and when hospitals are reimbursed predominantly by other payers (Models 2 and 4, p < 0.05). Hospital entry rates were significantly higher when diffuse facilities and diffuse hospitals receive a low proportion of reimbursements from private payers (Models 3 and 4, p < 0.01). As expected, both ASC and hospital entry rates had a positive relationship with the size of the population age 0–64 (Models 2 to 4, p < 0.05), but hospital entry rates had a negative relationship with the diffuse size of the population age 65+ (Model 3, p < 0.05). Across the board, ASCs tended to enter markets in which there were a relatively low proportion of minority residents (Models 1 to 4, p < 0.05). Per capita income also had the expected positive relationship with facility entry rates, although ASC entries were affected more strongly by local per capita income (Models 3 and 4, p < 0.01) while hospital entries were affected more strongly by diffuse income levels (Models 3 and 4, p < 0.05).

Regarding our hypotheses, we find strong support for hypothesis 1, which suggests that hospitals entry rates are negatively related to niche overlap density among outpatient facilities (Model 3, p < 0.01), which includes ASCs and hospitals (Model 4, p < 0.05). Our finding on the effect of overlap density on entry patterns within a mature organizational population of hospitals was as predicted.

We also find support for hypothesis 2 as our results indicate that surgery center entry rates are highest in markets with high niche overlap density among local and diffuse outpatient facilities (Models 1 and 3, p < 0.05) and particularly among surgery centers (Model 4, p < 0.05). However, surgery center entry rates do not appear to be influenced by hospitals overlap density. Taken together, these results provide support for the idea that firms within a growing organizational population tend to seek out crowded niches. This is the exact opposite phenomenon of what we observed from hospitals.

Discussion

The confluence of results from our models suggest that entry patterns do differ significantly within the mature population of hospitals and the growing population of ASCs. Hospitals tend to avoid direct competition with surgery centers and other hospitals by not entering markets with higher number of ASCs and hospitals with overlapping niches. Meanwhile, surgery centers appear to seek out markets in which there are surgery centers with overlapping niches. In simple words, at this stage of their evolution, ASCs are not hindered by the level of competition in the markets they plan to enter. While in the hospital population, competitive pressures discourages hospitals from entering markets with higher level of competition. Our findings also indicate that ASCs have higher entry rates in markets with higher demand, per capita income, and publicly and privately insured individuals, and lower proportion of minority groups. This is not surprising since the majority of ASCs are for-profits and entrepreneurs might view these markets as highly favorable.

Hospitals play a critical role in any healthcare system both in terms of the services they provide and also in terms of the workforce they employ. Previous research indicates that hospital exit rates are higher in markets characterized by competition from ambulatory surgery centers.² Our findings indicate that entry rate of hospitals into such markets is lower. Therefore, based on our findings on exit and entry rates, healthcare systems worldwide that are considering or already encouraged a more competitive healthcare delivery landscape should consider the impact market entry of specialized healthcare delivery organizations have on the existing hospital population. Moreover, these specialized healthcare organizations are targeting traditionally profitable service lines such as surgery and orthopedics.²³ For instance, previous research indicates that there is a relationship between the market entry of specialty hospitals and the operating margins of the procedures they specialize in.²⁴ Therefore, there are concerns that as these population grow, general hospitals’ ability to cross-subsidize less profitable service lines such as psychiatric services might diminish. Finally hospitals since their emergence have played a key role in their local communities. Their closure and redistribution would pose concerns in terms of patient access and continuity of care.

It is valuable to note that ASCs, and even specialty hospitals, represent just the tip of the iceberg as there has been similar growth among other specialized providers dedicated to gastrointestinal endoscopy, diagnostic imaging, sleep disorders, peripheral vascular disease, cosmetic surgery, radiation therapy, lithotripsy, cardiac catheterization, and cancer chemotherapy.²⁵ These facilities may have a competitive advantage when competing with general hospitals as a result of their leaner organizational form. While the changing landscape of healthcare delivery organizations might have substantial impact on general hospitals, many on the other hand argue that specialization leads to performance improvement and better patient outcomes. Christensen et al.²⁶ argue that focused healthcare organizations such specialty hospitals, ASCs, eye centers, and other specialized organizations are value adding process businesses that are capable of delivering comparable level of quality, if not better, at lower prices. This is supported by empirical research which found an association between better outcomes and specialization in healthcare delivery organizations.^27,28 Moreover, in a comparative study in the United Kingdom, specialized treatment centers were found to be more efficient as evident in their shorter length of stay.²⁹

Our findings have implications from a strategy perspective. Given that these specialty organizations compete with hospitals directly on their most profitable service lines,³⁰ hospitals should be aware that the market entry of specialized healthcare organizations will encourage more focused organizations to enter their market, even as the market gets crowded. Hospitals in more profitable markets with lower proportion of minority residents, higher size of age 0–64 population, and higher per capita income, need to especially be aware of this phenomenon. While this seems to indicate potential undesired competition for hospitals from specialized organizations, incumbent hospitals that are already in crowded markets need not fear the entry of other general hospitals since that event is unlikely. Thus, hospitals can focus their efforts on strengthening their position and competitiveness in the market against current competitors by improving their efficiency and quality.

In conclusion, in this paper, we provide a conceptual framework that can be applied to study the market entry in health services research that extends beyond surgery centers into other specialized providers that have caused a variety of services to migrate away from the general hospitals. It is important to note though that eventually the rapid growth exhibited by surgery centers should slow as the industry experiences a shakeout and ultimately reaches maturity. During this time, the r-selection processes we have examined will give way to K-selection processes. Given 10 more years of data and further industry evolution, it would be fascinating to study firm entry and exit patterns during this transitional period since we only examine them at opposite ends of the spectrum.

References

Van de Ven

Beck

Buchner

. Preconditions for efficiency and affordability in competitive healthcare markets: are they fulfilled in Belgium, Germany, Israel, the Netherlands and Switzerland? Health Policy 2013; 109: 226–245.

Al-Amin

Housman

. Ambulatory surgery center and general hospital competition: entry decisions and strategic choices. Health Care Manage Rev 2012; 37: 223–234.

Al-Amin

Zinn

Rosko

. Specialty hospitals market proliferation: strategic implications for general hospitals. Health Care Manage Rev 2010; 35: 294–301.

Iglehart

. The emergence of physician-owned specialty hospitals. N Engl J Med 2005; 352: 78–84.

Carey

Burgess

Young

. Specialty and full-service hospitals: a comparative cost analysis. Health Serv Res 2008; 43: 1869–1887.

Schneider

Ohsefedt

Morrisey

. Effects of specialty hospitals on the financial performance of general hospitals, 1997–2004. Inquiry 2007; 44: 321–344.

Bureau of Health Professions. Area resource file, US Department of Health and Human Services, Health Resources and Services Administration, 2008–2009, 2008–2009.

Winter

. Comparing the mix of patients in various outpatient surgery settings. Health Aff 2003; 22: 68–75.

Housman

. Organizational form and niche overlap: the dynamics of surgery center entry and exit, Unpublished Dissertation, 2009.

10.

Brittain

Freeman

Organizational proliferation and density-dependent selection. In: Kimberly

Miles

(eds). The organizational life cycle: issues in the creation, transformation, and decline of organizations, San Francisco, CA: Jossey-Bass Publishers, 1980, pp. 291–338.

11.

Sohn

. A relational approach to measuring competition among hospitals. Health Serv Res 2002; 37: 457–482.

12.

Mendel

Ruef

Scott

. An organizational field approach to resource environments in health care: comparing entries of hospitals and home health agencies in the San Francisco Bay region. Health Serv Res 1998; 32: 775–803.

13.

Abraham

Gaynort

Vogt

. Entry and competition in local hospital markets. J Ind Econ 2007; 55: 265–288.

14.

Baum

Singh

. Organizational niches and the dynamics of organizational mortality. Am J Sociology 1994; 100: 346–380.

15.

Baum

Singh

. Organizational niches and the dynamics of organizational founding. Organization Sci 1994; 5: 483–501.

16.

Baum

Haveman

. Love thy neighbor? Differentiation and agglomeration in the manhattan hotel industry, 1898–1990. Admin Sci Q 1997; 42: 304–338.

17.

Baum

Oliver

. Toward an institutional ecology of organizational founding. Acad Manage J 1996; 39: 1378–1427.

18.

Sorensen

. Recruitment-based competition between industries: a community ecology. Ind Corp Change 2004; 13: 149–170.

19.

Dobrev

Kim

T-Y

Hannan

. Dynamics of niche width and resource partitioning. Am J Sociology 2001; 106: 1299–1337.

20.

MacArthur

Wilson

. The theory of island biogeography 1967; Vol xi, Princeton, NJ: Princeton University Press, pp. 203–203.

21.

Makuc

Haglund

Ingram

. Health service areas for the United States. Vital Health Stat 2 19911121–102.

22.

Kessler

McClellan

. Is hospital competition socially wasteful? Q J Econ 2000; 115: 577–615.

23.

Mitchell

. Utilization changes following market entry by physician-owned specialty hospitals. Med Care Res Rev 2007; 64: 395–415.

24.

Schneider

Miller

Ohsfeldt

. The economics of specialty hospitals. Med Care Res Rev 2008; 65: 531–553.

25.

Berenson

Ginsburg

May

. Hospital-physicians relations: cooperation, competition, or separation? Health Aff 2007; 26: w31–w43.

26.

Christensen

Grossman

Hwang

. The innovator’s prescription: a distruptive solution for healthcare, New York, USA: McGraw-Hill, 2009.

27.

Chukmaitov

Devers

Harless

. Strategy, structure, and patient quality outcomes in ambulatory surgery centers (1997–2004). Med Care Res Rev 2011; 68: 202–225.

28.

Greenwald

Cromwell

Adamache

. Specialty versus community hospitals: referrals, quality, and community benefits. Health Aff 2006; 25: 106–118.

29.

Siciliani

Sivey

Street

. Differences in length of stay for hip replacement between public hospitals, specialised treatment centres and private providers: selection or efficiency? Health Econ 2013; 22: 234–242.

30.

Devers

Brewster

Casalino

. Changes in hospital competitive strategy: a new medical arms race? Health Serv Res 2003; 38: 447–469.