JOURNAL HOME CME HOME THIS MONTH PAST ISSUES ETOC COLLECTIONS
AUTHORS REVIEWERS EDITORIAL BOARD FEEDBACK RSS HELP
A&A International Anesthesia Research Society
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a colleague
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wachtel, R. E.
Right arrow Articles by Dexter, F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wachtel, R. E.
Right arrow Articles by Dexter, F.

Anesth Analg 2007;104:1157-1170
© 2007 International Anesthesia Research Society
doi: 10.1213/01.ane.0000260335.08877.3e

ECONOMICS, EDUCATION, AND POLICY

Section Editor:
Franklin Dexter

Application of a Similarity Index to State Discharge Abstract Data to Identify Opportunities for Growth of Surgical and Anesthesia Practices

Ruth E. Wachtel, PhD, MBA*, Elisabeth U. Dexter, MD, FACS{dagger}, and Franklin Dexter, MD, PhD*{ddagger}

From the *Division of Management Consulting, Department of Anesthesia, University of Iowa, Iowa City, Iowa 52242; {dagger}Division of Thoracic Oncology, Department of Surgery, State University of New York Upstate, Syracuse, New York 13210; and {ddagger}Department of Health Management and Policy, University of Iowa, Iowa City, Iowa 52242.

Address correspondence to Franklin Dexter, MD, PhD, Division of Management Consulting, Department of Anesthesia, University of Iowa, Iowa City, Iowa 52242. Address e-mail to Franklin-Dexter{at}UIowa.edu. For more information visit www.FranklinDexter.net

Abstract

INTRODUCTION: Most surgical and anesthesia groups are interested in expanding their practices and recruiting more patients. Methods have been developed to help hospitals identify surgical specialties with the potential for growth by determining whether the hospital is performing fewer of certain types of procedures than expected in a given specialty. However, these methods are not appropriate for physicians who may practice at more than one hospital and want to determine the potential for growth in their regions.

METHODS: We examined potential markets for growth of surgical and anesthesia practices in Iowa and New York State using state discharge abstract data. Several patient demographic groups and several surgical specialties were examined. Each state was divided into regions, and data were analyzed three ways: (1) A similarity index compared each region to the rest of the state. (2) The number of procedures performed on patients who left their home regions for care was determined. (3) A similarity index compared procedures performed on patients who left their home regions for care with procedures performed on patients who remained within their home regions.

RESULTS: The methods successfully identified several geographic regions with previously unrecognized growth potential. Access to care was limited in these regions. The methods correctly showed few opportunities for growth in geographic regions where expansion was already known to be unlikely.

CONCLUSIONS: A count of the number of procedures performed on patients who left their home regions, in combination with the similarity index, is a useful method for screening state discharge abstract data to identify geographic regions where surgical and anesthesia practices could grow.

Many surgical groups are interested in expanding their practices and recruiting more patients. Given the increasing competition in the health care industry, both surgical and anesthesia practices must be proactive in looking for growth opportunities. Nevertheless, anesthesia groups must be cautious. Hiring additional anesthesia providers in anticipation of surgical growth poses a financial risk because the number of cases may not increase as expected. If the anesthesia groups wait until the caseload has increased before they hire additional staff, long work hours and much frustration may ensue.

Individual hospitals can use state discharge data and internal financial data to assess opportunities for growth and provide information necessary for strategic planning. (a) Different hospitals in the state can be compared with respect to the numbers and types of surgical procedures they perform (1,2). (b) A hospital can identify other facilities that may be substantively affecting its caseload, and can perform a competition analysis to determine factors involved in patients’ decisions about where to go for surgery (3). (c) The financial implications of performing certain types of procedures can be assessed (4). (d) Data envelopment analysis (5–8) can determine which inpatient surgical specialties have the potential for growth in hospitals with >200 beds. By examining the market visibility of individual hospitals and estimating the need for medical services in the local area, data envelopment analysis determines whether a hospital could be performing more of certain types of procedures in a given surgical specialty based on its workload of procedures in other specialties. (e) If the hospital’s surgical workload is constrained by the number of cases it can readily accommodate each day, then operating room time can be allocated preferentially to those surgeons whose contribution margins are above average (8–11).

Methods developed for hospitals may not be appropriate for surgical and anesthesia groups. The hospitals at which they practice may be too small to permit meaningful comparisons between hospitals or between specialties within a single hospital. Surgical and anesthesia groups are not restricted to a single hospital, and may practice at more than one facility in a region. Recruiting patients away from one hospital and increasing caseload at another hospital where the same physicians also practice is of no value in promoting growth of the surgical and/or anesthesia group(s). Thus, analysis by region, rather than hospital, may be more useful for group practices.

This article describes the use of screening methods to identify geographic regions that have the potential for growth of surgical and anesthesia practices. The methods use information from state databases to identify patients who leave their home regions to receive care elsewhere. Groups within those regions may be able to increase their caseloads by persuading patients who need surgery to remain closer to home (12,13).

Results of earlier analyses of several demographic groups and surgical specialties were used to assess the validity of these new methods and confirm that they yield appropriate results.

METHODS

Specified operative procedures were studied in Iowa and New York State. Each state was first divided into regions. The analysis then involved three steps:

  1. A similarity index was used to compare the numbers and types of procedures performed in each region to the rest of the state. If the similarity between a particular region and the rest of the state was low, then the region was either (a) not offering needed services for which patients must go elsewhere, or (b) offering services not provided in other regions.
  2. A count was obtained of the number of procedures performed on patients who left each region to receive care in another region.
  3. A similarity index was used to compare procedures performed on patients who left each region for care with procedures performed on patients who remained within the region for care. If the similarity was low, then patients were leaving for care not provided in the region.

Definition of Regions
Counties in Iowa were divided into the seven regions defined by the Iowa Hospital Association (14) (Fig. 1). A pediatric hospital associated with an academic medical center is in Region G, while two other pediatric hospitals are in Region E.


Figure 129
View larger version (23K):
[in this window]
[in a new window]

  		Figure 1. Iowa regions as defined by the Iowa Hospital Association (14).

 

Counties in New York State were divided according to the areas served by the seven regional offices of the New York Department of Health (15) (Fig. 2A). To determine the sensitivity of the analyses to the choice of regional boundaries, regions were also defined according to the 8 New York State Statewide Planning and Research Cooperative System (SPARCS) Health Service Areas (17), the 11 Department of Transportation regions (18), and the 11 "Visit New York" tourist regions (19).


Figure 229
View larger version (14K):
[in this window]
[in a new window]

  		Figure 2. A. New York regions as defined by the New York State Department of Health (15). 1. Capital District, 2. Central New York, 3. Western Region—-Buffalo, 4. Western Region—-Rochester, 5. New York Metropolitan—-New Rochelle, 6. New York Metropolitan—-Long Island, 7. New York Metropolitan—-New York City. B. Map of major roads in New York (16) and the physical boundaries that prolong travel between adjacent regions, such as Lake Ontario to the north and the Adirondack Mountains in the northeast. The major roads run in a north-south direction within each region, and few roads run between regions. Geographical regions are defined to match naturally occurring barriers to travel.

 

Similarity Index
The similarity index is a correlation coefficient between two regions or groups of procedures that describes the degree of overlap based on the relative numbers of the different types of procedures performed. It was used (1) to compare the relative frequencies at which different types of procedures were performed in a specific region compared with those performed in the rest of the state. The similarity index was also used to compare procedures performed on patients who left a region to go elsewhere for care with procedures performed on patients who chose to stay within their home regions.

Suppose we want to compare the proportion of procedures performed between two groups of patients. Let p1k and p2k represent the proportion of procedures performed in Groups 1 and 2 that are of the kth type of procedure, k = 1, 2, ..., T, where T is the total number of different types of procedures. For example, Group 1 could be procedures performed in any of the seven different regions in one state, and Group 2 would then be procedures performed in the entire rest of the state. Alternatively, Group 1 might be procedures performed on patients in a certain age category who left one of the seven defined geographical regions for care, while Group 2 would then be procedures performed on patients in the same age category who remained in that particular region.

The similarity of the relative frequencies with which the different types of procedures are performed is (1):



Formula 1

If two procedures are selected at random, one from each group or region, each term in the numerator is the probability that both procedures will be of the same type. The denominator normalizes the sum of the probabilities to a value between 0 and 1. The similarity index equals zero when there is no overlap in the types of procedures performed and it equals one when the relative frequencies are the same for both groups for all types of procedures. Cramér’s delta method was used to obtain the standard error for {theta}, as was validated previously (1).

Published values comparing groups of surgical procedures have reported similarities of approximately 0.6–0.8 (1). These intermediate values are difficult to interpret. For the similarity index to be a useful screening tool, it would have to exhibit a much wider range of values for the application being studied. Values of 0.3 or less would clearly indicate meaningful differences between groups. Values of 0.9 or higher would provide unambiguous evidence that geographic areas or surgical practices are very much alike.

Patients Who Left Each Region
To assess the potential for growth of a surgical practice within a given region, we counted the number of procedures performed on patients living in each region who traveled outside that region for surgical care. Confidence intervals (95% CI) for the number of procedures performed on patients leaving a region each week for surgery were calculated assuming a Poisson distribution (20). As above, the similarity index was calculated to assess the extent of overlap between the numbers and types of procedures performed on patients who stayed in their home regions and those performed on patients who left each region to receive care elsewhere.

Databases
Data were taken from the Iowa inpatient and outpatient discharge abstract databases, January 1, 2001, to June 30, 2001. Pediatric surgery was studied in infants and young children 0–2 years of age (1), and geriatric surgery was studied in patients 80 years of age and older (2). These dates and ages were chosen to match those of prior analyses (1–3,5). The discharge abstract databases included cases performed in every nonfederal hospital and hospital-affiliated outpatient surgery center statewide. Procedures and diagnoses were coded by each facility using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM).

Thoracic procedures in New York State were obtained from the SPARCS (21) database for 2002. The database included ICD-9-CM procedure codes and patient and facility information for surgical procedures performed at all inpatient and ambulatory facilities statewide.

Operative Procedures
We studied only "operative procedures," defined as those associated with an operating room and/or anesthesia charge, and meeting the additional requirement that an incision be made (22). For example, cardiopulmonary bypass (extracorpeal circulation, heart–lung machine; ICD-9-CM 39.61) was excluded because the code refers only to operating the machine. The incision and preparation for bypass are considered part of the associated surgical procedure(s). Diagnostic procedures (ICD-9-CM 87.0 and greater) were also excluded (e.g. magnetic resonance imaging).

Physiological Complexity of Procedures
A procedure was considered physiologically complex if it had 8 or more basic units according to the 1999 American Society of Anesthesiologists’ Relative Value Guide (23–25). Laparoscopic cholecystectomy (7 units) is not physiologically complex, while hip replacement (8 units), carotid artery bypass (10 units), lung lobectomy (13 units), and abdominal aortic aneurysm resection (15 units) are all physiologically complex.

Thoracic Procedures
Procedures typically performed by general thoracic surgeons were analyzed to determine the potential for growth of general thoracic surgery. A list of 27 different types of procedures for study was derived from the 2004 Clinical Classifications Software categories 36 and 42 (Table 1), where Clinical Classifications Software (26) is the Agency for Healthcare Research and Quality collapse of the 3500 ICD-9-CM procedure codes into a smaller number of clinically meaningful categories. For a thoracic procedure to be included in this study, the primary surgical procedure had to be one of the specified types of procedures. Procedures performed ancillary to other types of surgery were therefore excluded, such as a lung biopsy during a cardiac operation. Patients 12 years of age and younger were excluded to eliminate overlap with the specialty of "pediatric surgery."


View this table:
[in this window]
[in a new window]

  		Table 1. ICD-9-CM Used to Define Thoracic Procedures

 

RESULTS

Four demographic groups were used to validate the three techniques. The techniques are: (1) using a similarity index to compare each region to the rest of the state, (2) counting the numbers of patients leaving each region, and (3) using a similarity index to compare procedures between patients who stayed and those who left each region.

Although the first technique was of limited usefulness in certain situations, the three methods successfully identified several regions with growth potential that earlier studies had failed to recognize (5). The methods also demonstrated few opportunities for growth in geographic regions where expansion was already considered unlikely (1,2,5).

Infants and Young Children in Iowa
Background
The distribution of surgical procedures among Iowa hospitals has been examined previously for infants and young children 0–2 years of age (1,2). Region G contains the state’s largest pediatric hospital, also called "the academic medical center," which dominates surgery in this age group. It performed 64% of all physiologically complex surgery and 63% of all rare physiologically complex surgery in the state (1,2). Of the 246 different types of procedures done statewide, 181 (86%) were performed at the academic medical center (1). Region E contains a smaller pediatric hospital that performed 73 different types of procedures and 10% of all physiologically complex surgery in 0–2-yr-olds (1). Another pediatric hospital in Region E performed 58 different types of procedures and 24% of all physiologically complex surgery in 0–2-yr-olds (1). In contrast, the highest volume hospital in Iowa for infants and young children performed only seven different types of procedures, 99% of which were myringotomy tube placement and tonsillectomy and/or adenoidectomy (1,2).

Predictions
These earlier findings suggested to us that the two regions containing pediatric hospitals should have little opportunity for growth. In contrast, opportunities for expansion of pediatric surgery should be present in the other regions.

  1. The two regions containing pediatric hospitals provide specialized services, such as physiologically complex surgery, that are not offered elsewhere. The similarity between each of these regions and the rest of the state should therefore be low. Each of the regions not containing a pediatric hospital should be more representative of the state as a whole, with a higher similarity index.
  2. The number of patients leaving the two regions containing the pediatric hospitals should be small. A substantial number of pediatric patients should leave the other regions for care, especially for physiologically complex surgery involving inpatient care.
  3. The similarity should be low between procedures performed on infants and young children who leave geographical regions not containing a pediatric hospital and procedures performed on patients who remain within those regions. In other words, patients leave for types of procedures not performed within their home regions.

Findings

  1. Contrary to our prediction, the similarity index was high (≥0.94) between each of the 7 Iowa regions and the rest of the state (Table 2, Column C). The two regions (E and G) that are known to differ from the rest of the state due to the presence of pediatric hospitals (1) did not exhibit lower similarity indices. The similarity index may thus be unable to identify differences that are present and may give rise to false-negative results. When each region is compared to the rest of the state, the similarity index may fail to identify regions that are known to be different.

    Previous studies comparing surgical procedures among hospitals and/or regions yielded only intermediate values in the range 0.6–0.8 (1,2). This is the first report showing that the similarity index can be so high.
  2. As predicted, very few patients left Region E for care. A relatively large number of procedures were performed on patients who left Regions B and F to travel to other regions (Table 2, Column E). If patients undergoing these procedures had remained within Regions B and F, each hospital within the region could have increased its workload for surgery in 0–2-yr-olds by 32%–33% (Column F). Thus, the number of procedures performed on patients who left each geographical area provides useful information for identifying opportunities for growth of surgical and anesthesia practices.
  3. When the similarity index was used to compare procedures performed on patients who left Regions B and F with those performed on patients who stayed within each region, similarities were 0.84 and 0.93 (Table 2, Column G). These values are quite high, indicating that the types of procedures that left are not substantially different from those that stayed within each region. Figure 3A shows a graphical representation of the information used to calculate the similarity index for Region F. It illustrates the variety and relative proportions of the different types of procedures that left Region F. Each circle represents one type of procedure. Many types of procedures were not performed within the region, but were performed once or twice on patients who left the region. For the larger circles that represent greater numbers of cases, however, the relative frequencies at which the various types of procedures were performed is similar for patients who stayed in the region and those who left the region. In contrast to predictions, patients often leave for types of procedures already being performed within their home region. Region F has the potential to increase its caseload if it can capture some of those patients who are leaving the region for types of procedures it is already doing.


View this table:
[in this window]
[in a new window]

  		Table 2. Surgical Procedures Performed on Infants and Young Children in Iowa During a 6-mo Time Period
 

Figure 329
View larger version (11K):
[in this window]
[in a new window]

  		Figure 3. A. For 0–2-yr-olds, comparison between procedures performed on patients who left Region F and those performed on patients who stayed in the region. Each type of procedure is shown as a separate point. Its position along each axis is determined by the fraction of all procedures that are due to that type of procedure. The size of each point is related to the number of times that type of procedure was performed. The largest points thus represent those types of procedure that constitute a much higher proportion of procedures than any other type, and are performed more frequently than other types of procedures. Points with zero as one of their coordinates have been offset slightly to avoid concealing overlapping points. Numerous types of procedures not performed within the region were performed a single time outside the region. When the size of each point is taken into consideration, the types of procedures that left the region and the types of procedures that stayed within the region were generally performed in similar proportions. B. Comparison between procedures that stayed in Region B and those that left the region, with the exclusion of ICD-9-CM 20.01 (placement of myringotomy tube). Each type of procedure that stayed in the region is shown as a shaded bar, with the height of the bar proportional to the number of that type of procedure that stayed in the region. White bars with a dark outline show the number of each type of procedure that left the region. Many types of procedures that left the region occurred only once and were not performed at all within the region. Procedures are sorted in decreasing order of frequency.
 
Infants and Young Children in Iowa after Exclusion of Myringotomies
Similarity indices may have failed to identify differences between regions containing a pediatric hospital and those without a pediatric hospital (method [1]), and differences between procedures performed on patients who remained in their home regions and those who left (method [3]), because one type of procedure accounted for 73% of all procedures (ICD-9-CM 20.01 placement of myringotomy tube). Myringotomies can be seen in Figure 3A as the largest point, because they represent the type of procedure performed more frequently than any other type. This common procedure may have obscured other types of procedures that required specialized care at a pediatric hospital. Table 3 therefore shows analyses of the same pediatric data as Table 2, except for the exclusion of this one type of procedure.


View this table:
[in this window]
[in a new window]

  		Table 3. Analyses of the Same Data Used for Table 2, with the Exclusion of Myringotomy Tube Placement

 

  1. After exclusion of myringotomy tube placement, the similarities between each region and the rest of the state (Table 3, Column C) are now 0.5–0.8. When Regions E and G are compared to the rest of the state, however, the similarities are still not lower than those for the other regions. The similarity index is not a sensitive tool for distinguishing between regions or identifying regions that are different.
    The simplest explanation for this lack of sensitivity is that pediatric surgery in Iowa represents a unique situation. The numbers and types of pediatric procedures performed at the academic medical center in Region G do not differ substantially from those performed by the two pediatric hospitals combined in Region E. When Regions G and E are compared to each other, the similarity index is 0.996 ± 0.001, or 0.88 ± 0.03 after exclusion of myringotomy tube placement. Thus, each region containing a pediatric hospital is similar to the rest of the state because the "rest of the state" includes another region whose pediatric hospital(s) performs the same numbers and types of procedures as the region of interest. The similarity index, when used to compare regions to the rest of the state, does not identify outlier regions because there are two outlier regions that are similar to each other.
    Individual pediatric hospitals were then used as positive controls to ensure that the similarity index was being applied correctly. When each of the pediatric hospitals was compared to the rest of the state, the similarities were low, as expected (bottom of Table 3). Although the academic medical center in Region G performed the majority of procedures within Region G, the hospital is much less similar to the rest of the state than Region G is to the rest of the state. The academic medical center performed more rare procedures and relatively fewer tonsillectomies and/or adenoidectomies than did other hospitals in the region. The similarity between the academic medical center and the rest of Region G is only 0.10 ± 0.02. Comparisons involving individual hospitals demonstrate that the similarity index appropriately shows differences when differences are known to exist. The absence of significant differences when comparing regions is not an artifact due to improper implementation of the technique, but results rather from the unique situation of having two regions that are similar to each other.
    Use of individual hospitals as control groups also demonstrates that the similarity index can assume extremely low values, such as 0.10 when comparing the academic medical center to the rest of Region G, or 0.14 when comparing it to the rest of the state (last row of Table 3). Such low values have not been reported previously when comparing hospitals and/or surgical procedures.
  2. As predicted, few patients left the regions with pediatric hospitals (E and G) for care, equivalent to one single procedure per week (Table 3, Column E). Because of the large number of procedures performed in these two regions (Column B), the percentage increase in caseload would be negligible even if all patients who left these two regions remained there for care (Column F). In contrast, many patients left Regions B and F for care relative to the number who remained. A simple count of the number of procedures performed on patients who left each region shows that opportunities for growth are limited within Regions E and G, but may be substantial within regions B and F.
  3. After excluding myringotomy tube placement, Regions B and C are noteworthy because substantial numbers of patients left these regions (Column E), and the similarity indices are low when comparing procedures performed on patients who left with those performed on patients who remained (Table 3, Column G). As predicted, the similarity index demonstrates that patients left those regions for specialized care and certain types of procedures not performed within the regions. Surgical and anesthesia practices in these regions have an opportunity to increase their caseload by recruiting patients who leave these regions and inducing them to have surgery closer to home. However, the surgeons will have to expand the scope of their practices beyond the types of procedures presently offered.

The low similarity index (0.35) comparing procedures performed on patients who left Region B and those who stayed in the region is illustrated in Figure 3B. Each type of procedure is shown as a separate bar. Many types of procedures that left the region were performed only a single time on patients from the region, consistent with the low similarity index and the conclusion that patients left to undergo rare or specialized types of procedures. Region B has the potential to expand its caseload, assuming it has the capability to perform a wider variety of procedures, each at a low frequency.

The conclusion that there are multiple opportunities for expansion of pediatric surgery and improved access to care in regions that do not contain pediatric hospitals is supported by data on inpatient procedures. Physiologically complex procedures (1,2,25) or serious comorbidities should be associated with inpatient stays. Infants and young children undergoing inpatient surgery would thus be more likely to have their surgery at hospitals that specialize in pediatric care. As predicted, the absolute number of inpatient procedures was much higher in the two regions with pediatric hospitals (E and G) than elsewhere (Table 4, Column B). Furthermore, with the exception of Region G containing the academic medical center, the fraction of procedures that were inpatient procedures was much higher in the group that left each region (Column D) than the group that stayed within each region (Column C). This finding is consistent with the interpretation that patients leave each region for specialized procedures not performed within the region. Patients would likely remain within their home regions if physicians there could provide specialized pediatric care.


View this table:
[in this window]
[in a new window]

  		Table 4. Inpatient Surgical Procedures on Infants and Young Children During a 6-mo Time Period

 

Physiologically Complex Surgery in Geriatric Patients in Iowa
Background
For physiologically complex surgical procedures in geriatric patients 80 years and older (2), surgery occurred at multiple hospitals distributed throughout Iowa. No one or two hospitals accounted for the majority of the 130 different types of procedures. Many hospitals performed procedures that others did not. When the academic medical center was compared to other hospitals, however, a larger percentage of the procedures it performed were of types considered rare (2). In addition, a greater percentage of patients at the academic medical center had traveled from outside their home or contiguous counties to come to that hospital relative to patients at other hospitals (2).

Predictions

  1. Because no one hospital dominated geriatric surgery (2), the similarity index comparing each region to the rest of the state should show that all regions are comparable. The academic medical center should be different from the rest of the state, however, because many of its procedures are of types that are rare (2).
  2. Small numbers of patients should leave the region of the academic medical center to travel to other regions for care. Some patients should leave other regions, however, to travel to the region of the academic medical center.
  3. Because most patients have surgery near their home counties, but some patients travel for rare procedures, the similarity between procedures performed on patients who leave each region and those who remain should be intermediate. In general, results will show limited potential for expansion of geriatric surgery.

Findings
Results were consistent with all predictions, except that Region F was unexpectedly different because of the low number of procedures performed.

  1. Most values of the similarity index are intermediate, ranging from 0.6 to 0.8 (Table 5, Column C). However, the similarity index comparing Region F to the rest of the state is only 0.39. As expected, the similarity between procedures performed by the surgical group at the academic medical center in Region G and those performed in the rest of the state is low (0.31, last row of Table 5, Column C).

  2. Relatively few procedures were performed on patients who left the region of the academic medical center for other regions (2.6 per week, Table 5, Column E), representing a potential increase in volume of only 11% (Column F). Patients left other regions to travel to the region of the academic medical center, however. About 34% of the procedures performed there came from outside the region (Table 5, the difference between Columns B and D divided by Column B).
  3. The similarity between procedures performed on patients who left each region and procedures performed on patients who remained within each region is intermediate (Table 5, Column G). An exception is Region F, which is apparently under-served.


View this table:
[in this window]
[in a new window]

  		Table 5. Physiologically Complex Surgical Procedures on Geriatric Patients in Iowa During a 6-mo Time Period
 
Analysis by region was extremely useful for identifying opportunities for growth in Region F. (1) On average, 6.7 procedures were performed each week on patients who left the region for care (Table 5, Column E), representing an increase in workload of approximately one case per day. (2) However, the region has a relatively low similarity index of 0.39 when compared to the rest of the state (Column C). (3) The similarity is even lower (0.18) when comparing procedures that left the region with those that stayed (Column G), indicating that patients are leaving for specialized procedures not performed within the region. Surgeons thus have an opportunity to perform additional cases if they can expand their scope of practice beyond those types of procedures already performed in the region.

An alternative interpretation accounting for the low similarity between Region F and the rest of the state is that Region F has created a niche in geriatric surgery and is unique in performing specialty surgical procedures not done elsewhere. This explanation is unlikely, mainly because Region F performed the fewest number of procedures of any region.

General Thoracic Surgery in Iowa
Background
Thoracic surgery is performed throughout the state, not just at specialized centers (4,5). Nevertheless, the academic medical center performed almost 25% of all physiologically complex thoracic procedures done statewide (4). Data envelopment analysis previously showed that the hospital had minimal potential for growth in this area based on local patient demographics (5).

Predictions

  1. Because thoracic surgery is performed throughout the state, similarities should be high between each region and the rest of the state.
  2. The number of patients leaving the region of the academic medical center for care in other regions should be small.

Findings

  1. As predicted, each region is highly similar to the rest of the state (Table 6, Column C), with the exception of Region F.

  2. Few patients left Region G, which contains the academic medical center (Table 6, Column E).


View this table:
[in this window]
[in a new window]

  		Table 6. General Surgery Thoracic Procedures in Iowa During a 1-yr Time Period
 
The usefulness of evaluating the number of procedures performed on patients leaving a region is supported by findings for Region F. There may be opportunities in Region F to improve access to care by expanding thoracic surgery services beyond current workload, though the actual number of additional procedures that might be performed would not be large.

General Thoracic Surgery in New York State
Background
New York State has two unusual geographical characteristics that distinguish it from Iowa: natural barriers to travel and a large metropolitan area. In upstate New York, barriers such as Lake Ontario, the Finger Lakes, and the Adirondack and Catskill Mountains limit the extent of travel in easterly and westerly directions (Fig. 2B). Travel across the northwest boundaries of the state is limited by Lakes Erie and Ontario, and by the Canadian border. Patients living in certain parts of the state must therefore travel long distances if they desire care in other regions. For example, residents of the border towns of Watertown or Burlington must first travel 50–100 miles south, or go north through Canada, before they can go east or west. In contrast, in the New York City metropolitan area, travel by car or rail is not restricted by natural barriers. Residents of Long Island would not likely travel upstate, however, because many large medical centers in New York City are much closer.

Predictions

  1. Because travel is restricted, thoracic surgery cannot be regionalized to a few hospitals within the state. All regions perform the full spectrum of thoracic procedures, and no region is unique. All regions within New York State should be similar to the rest of the state with respect to thoracic procedures.
  2. A count of the number of patients leaving each region should show that, in northern New York State, most patients stay in their home regions for care. Few patients should travel across natural barriers to other regions. In contrast, many patients should travel freely between Manhattan and Long Island in the New York City metropolitan area.
  3. The similarity between procedures performed on patients who left each region and those who remained within the region should be intermediate, around 0.6–0.8. This pattern will be true for both upstate New York, where few patients leave each region for care, and the metropolitan area, where many patients leave each region to travel to another region within the New York City area. There are few opportunities for expansion of thoracic surgery practices within the State of New York.

Findings

  1. As predicted, similarities are extremely high (Table 7, Column C) between each region and the rest of the state. Similarities are high whether the state is divided according to the 7 regions of the New York State Department of Health, the 8 regions of the SPARCS Health Service Areas, the 11 regions of the Department of Transportation, or the 11 regions of "Visit New York" tourism (data not shown). These findings illustrate both the robustness of the similarity index for comparing procedures and the validity of dividing the state into regions. Results are not sensitive to the specific definitions used to differentiate regions.

  2. In upstate New York (Regions 1–4), relatively few patients left each region for care, presumably because of geographical barriers to travel (Table 7, Column E; Fig. 2B). Similarities between procedures performed on patients who did leave each region and those who stayed in the region are intermediate to high (0.70–0.87, Column G), suggesting that patients are not leaving for specialized procedures not performed within the region. Each region in upstate New York therefore provided the full spectrum of services to meet the needs of its residents. No evidence was found to suggest substantive opportunities for growth of thoracic surgery practices in upstate New York. Results are consistent with a study of comprehensive pediatric hospitals, which showed that the population of New York State was served widely and evenly via eight referral regions (27).
    Findings relevant to the academic thoracic surgery and anesthesia groups in Syracuse illustrate the validity of counting the number of patients leaving each region. Excluding Veterans Administration patients, these academic groups draw thoracic patients almost exclusively from a 13-county service area (28). This service area coincides almost exactly with Region 2. Almost all procedures (97%) performed in Region 2 came from within the region (Table 7, Columns B and D). Few patients living in Region 2 left the region for care elsewhere (Column E).
    In dramatic contrast, many patients traveled for care within the New York City metropolitan area, Regions 5, 6, and 7 (Column E). However, 99% of the patients who left Region 6 (Long Island) went to Region 7 (New York City), and 96% of the patients who left Region 7 sought care in Region 6. Of patients who left Region 5, 88% traveled to Region 7. In all likelihood, those patients who left Region 5 to visit Region 7 went to a hospital that was closer to their residence than many hospitals in their own region. Few patients traveled from the metropolitan area to go upstate. If a surgical group in the metropolitan area wishes to increase its caseload, it must attract patients away from other hospitals and surgical practices that also serve the greater metropolitan area.
  3. As predicted, for both upstate New York and the New York City metropolitan area, similarities between procedures performed on patients living in a region and those performed on patients who left each region are intermediate to high (Column G). Patients did not leave a region for specialized procedures that were not performed within the region. No region was under-served.


View this table:
[in this window]
[in a new window]

  		Table 7. General Surgery Thoracic Procedures in the State of New York During a 1-yr Time Period
 
DISCUSSION

This paper describes the use of three new methods that surgical and anesthesia groups can use to identify geographic regions that have the potential for growth. Methods use state discharge abstract databases, and involve (1) applying a similarity index to compare individual regions to the rest of the state, (2) counting the number of procedures performed on patients leaving a region to undergo surgery elsewhere, and (3) using the similarity index to compare the types of procedures performed on patients who leave the region to those performed on patients who stay. These methods identify regions that may suffer due to a maldistribution of specialists. Residents must travel elsewhere for medical services, to locations where specialist care is more readily available.

While previous techniques have focused on individual hospitals (1–11), this study involves large regions within a state. Analysis by region takes less time because the number of regions is fewer than the number of hospitals. Analysis by region is also more appropriate for medical groups practicing at more than one hospital in a region.

(1) A similarity index can be applied to state discharge data and used as a screening tool for comparing each region to the rest of the state. The similarity index should identify regions that have a unique case mix and perform various types of procedures in proportions that differ from the rest of the state. A low similarity index would indicate that a region is already performing specialized procedures not done elsewhere and has little opportunity to expand, or that a region performs only a limited number of different types of procedures and represents a good opportunity for growth. In New York State, for example, the similarity between each region and the rest of the state is high. Each region is apparently able to meet the needs of its residents for thoracic surgery. The numbers of patients who leave the state for care, however, perhaps traveling south from Binghamton into Pennsylvania, cannot be estimated using discharge data from New York State. Nevertheless, the data clearly show that there are few opportunities for growth of general thoracic surgery by recruiting patients from other regions within the state.

In Iowa, however, the usefulness of the similarity index for comparing regions to the rest of the state is more limited. Results must be interpreted with caution. For pediatric surgery, similarity indices are high even though certain regions are known to differ from the rest of the state. Use of the similarity index as a screening tool failed to identify regions that are unique.

Other methods, when used in combination with the similarity index, are able to identify opportunities for growth. Methods involve (2) counting the number of procedures performed on patients leaving a region to undergo surgery elsewhere, then (3) using the similarity index to compare the types of procedures performed on patients who leave the region to those performed on patients who stay. The similarity index provides information on whether patients are leaving for the same types of procedures already performed in the region or for procedures not currently performed within the region. These methods can help a surgical or anesthesia practice identify opportunities for growth by demonstrating that the scope of care may be limited in certain areas. The methods also help to identify geographic areas with little growth potential.

Results show that the similarity index can span its entire range and assume values very close to its minimum of 0.0 and its maximum of 1.0. Such a wide range of values for the similarity index have not been reported previously. Earlier studies yielded intermediate values of 0.6–0.8 that were difficult to interpret (1). These new findings demonstrate the usefulness of the similarity index for providing convincing evidence that various demographic groups are clearly different or clearly similar.

Region F in Iowa serves as an excellent example of how these methods can identify an under-served region. If many patients are leaving a region for care, or if they are leaving for procedures not performed within the region, then the area may be under-served. Pediatric, geriatric, and thoracic patients all traveled to receive care not available within Region F. Region F does not appear unique with respect to patient demographics (29), the number of residents in poverty (30) or without health insurance (31), or the presence of small hospitals (32). Perhaps its close proximity to major medical centers in Regions E and G offers a possible explanation for the large numbers of patients who leave the region for care. Patients did not seek care at the major medical centers because the centers were close to their work location, however. Although patients often receive care near their work rather than their home (8), <3% of the workforce living in Region F works in the counties of the major medical centers (33).

Given patients’ reluctance to travel (12,13), surgeons in Region F may be able to increase their caseload by persuading patients to remain closer to home. Surgical practices may decide to improve access to care locally by hiring another surgeon with additional expertise in certain procedures, or by expanding the repertoire of procedures they routinely perform (34,35). Other surgeons could increase their caseload by moving to those regions that are under-served or traveling there several days a month to perform surgery as part of an outreach program (36). If patients who currently leave the region are already traveling to see those same surgeons who open an outreach clinic, the surgeons would see little increase in caseload. The patients would benefit, though, from improved access to care and greater convenience. Also, the finding that many patients travel for care within the state suggests that additional patients not included in this analysis could be seeing physicians outside the state. Surgeons and anesthesiologists could then increase their caseload by opening outreach clinics in the under-served region.

Hospitals in Regions G and E should realize that they are at risk if physicians in Region F, or medical practices from other areas, decide to improve access to care in Region F. Almost all patients who left Region F went to Regions G and E for care, constituting 77% and 22% of pediatric, 61% and 39% of geriatric, and 81% and 19% of thoracic procedures that left the region. Thus, hospitals in Regions G and E would see a decrease in caseload if patients in Region F were able to receive care within their home region.

For infants and young children, an average of 10 procedures a week were performed on patients leaving Regions B and F (Table 2, Column E). A surgical group in Region F could conceivably expand its surgical practice by hiring the equivalent of roughly 1–2 additional surgeons to care for those patients who currently leave the region. Within these regions, the highest volume surgeon averaged 8.2 procedures per week in 0–2-yr-olds. A typical surgeon who treats 0–2-yr-olds will total many more procedures than this because the surgeon will operate on older children as well as infants. The 10 procedures that left Regions B and F represent only that fraction of a surgeon’s practice that includes 0–2-yr-olds. Before deciding whether or not to expand, a hospital or group must therefore consider whether the number of procedures in other categories will also increase appropriately to account for the rest of the surgeon’s time.

If a surgical practice decides to expand by performing types of procedures not currently performed in the region, it must consider the resources available. Surgeons may not currently be skilled in the types of procedures to be performed. Anesthesia personnel may not have experience caring for unstable patients or small babies. The hospital may not have specialized equipment necessary to perform certain procedures, such as an operating microscope or robotic arm. Floor nurses may not have the expertise needed to care for patients postoperatively. Therefore, encouraging patients who leave a region for specialized procedures to remain within the region instead and receive care closer to home may not be practical or economical. In addition, it may not serve the best interests of the patient or the hospital.

Outcomes are correlated with surgical volume, especially for high-risk procedures such as major cancer surgery. This finding does not mean that expansion of surgical and anesthesia services into under-served regions will result in poorer outcomes. Local hospitals can perform certain specialized procedures. Access to care can be maintained or enhanced at lower volume hospitals by ensuring that surgeons are high volume (37,38) and that hospitals adopt appropriate patient care guidelines analogous to those used by high volume hospitals (38). Identification of potential regions for expansion of available services is a goal of the Medicare Rural Hospital Flexibility Program, which established a new designation for limited-service hospitals called critical access hospitals. The program, together with the Rural Community Hospital Assistance Act, provides financial assistance to smaller hospitals in an effort to prevent them from closing, thus ensuring continued access to care for rural residents (39).

Patients should not be forced to travel to regionalized centers for specialized care. Many patients live long distances from a high volume hospital (40,41) or experience barriers to travel (6), and thus referral to high volume centers is not always practical (40,42). In addition, regionalization would represent a change in current practices. The vast majority of hospitals performing high risk surgeries are not considered high-volume hospitals for those procedures (43). Most patients still have high risk surgery in low volume hospitals (40,41,43). For example, three-fourths of surgical procedures for esophageal cancer are performed at hospitals that do fewer than seven such procedures each year (43). Almost two-thirds of surgeries for cerebral aneurysm repair are done at hospitals that do fewer than 30 per year (43).

LIMITATIONS

We have shown that the usefulness of the similarity index for comparing each region to the rest of the state is limited when two regions are similar to each other, but different from the rest of the state. The similarity index as a screening tool may fail to identify regions that are unique.

The Iowa outpatient database contained data from hospital-affiliated surgical centers. Procedures performed at free-standing physician-owned ambulatory surgery centers were not included. The number of myringotomies, and other relatively minor procedures performed in children, may therefore have been under-estimated. No other demographic groups that we studied would have been affected, because the types of procedures involved would not be performed on an outpatient basis. If the number of pediatric cases were higher than reported here, such results would merely strengthen our conclusion that the similarity index was not sufficiently sensitive to detect differences between individual regions and the rest of the state.

The three methods studied may not be appropriate for a state in which large numbers of patients come from other states or leave the state to seek care elsewhere. For Iowa, only 6% of the 360,000 inpatient discharges in the 2004 database came from outside the state. Of that 6%, 92% of discharges involved patients from the six states that border Iowa. However, we do not know how many patients left Iowa to travel to other states. In Region F, which was under-served, only 1% of the workforce was employed in the neighboring states of Missouri and Illinois. (33) Thus, in all likelihood, residents of Region F did not travel to those states for surgery. For a state bordering large metropolitan areas, such as New Jersey, many patients may travel out of state for care (e.g., to New York City or Philadelphia). Thus, a count of the number of patients leaving each region could significantly under-estimate the true number of patients leaving the region. Patient travel from one region to another within the state may be unimportant relative to patients who leave the state.

For states in which large numbers of patients do not live within the state, comparisons of the types of procedures performed within different regions would still be valid. However, data would not be available concerning the number of potential patients who could have traveled to the state for care, but went elsewhere instead.

We did not examine local area variation in surgical rates. No data were available on the number of additional surgical procedures that could potentially be performed in a region if access to surgical care were more convenient, surgeons advertised the availability of elective procedures such as cosmetic surgery, referral patterns from internists were different, surgeons were more aggressive in pursing treatment, patients were given fewer alternative treatment options, insurance coverage were more widely available, or the prevalence of health maintenance insurance plans were different.

Surgical and anesthesia practices and hospitals were treated as "commodities" (3,44). We assumed that, except for location, medical services offered by one group were indistinguishable from the same services offered by another group. Issues such as reputation and market visibility were not considered. Factors other than geography that may have influenced patient choices about where to seek care are not addressed by the techniques described in this paper.

Although we concluded that Region F was under-served because so many patients left to receive care elsewhere, alternative explanations cannot be excluded. Perhaps physicians within Region F referred patients for surgery more frequently than physicians in other regions. Most residents of Region F would be close to major medical centers in Regions E and G, and residents of Region F may have been more inclined to travel for major surgery than residents of other regions. Residents of other regions may have chosen other therapeutic options because they did not wish to travel long distances.

CONCLUSIONS

Three new methods have been described for helping surgical or anesthesia practices identify geographic regions that have the potential for growth through an increase in the number of surgical procedures retained within those regions.

Footnotes

Accepted for publication January 5, 2007.

Dr. Franklin Dexter, Section Editor for Economics, Education and Policy, was recused from all editorial decisions related to this manuscript.

FD is Director of the Division of Management Consulting within the Department of Anesthesia. He receives no funds personally other than his salary from the State of Iowa, including no travel expenses or honoraria, and has tenure with no incentive program.

REFERENCES

  1. Dexter F, Wachtel RE, Yue JC. Use of discharge abstract databases to differentiate among pediatric hospitals based on operative procedures: surgery in infants and young children in the state of Iowa. Anesthesiology 2003;99:480–7.[ISI][Medline]
  2. Wachtel RE, Dexter F. Differentiating among hospitals performing physiologically complex operative procedures in the elderly. Anesthesiology 2004;100:1552–61.[ISI][Medline]
  3. Dexter F, Wachtel RE, Sohn MW, et al. Quantifying effect of a hospital’s caseload for a surgical specialty on that of another hospital using multi-attribute market segments. Health Care Manag Sci 2005;8:121–31.[Medline]
  4. Wachtel RE, Dexter F, Lubarsky DA. Financial implications of a hospital’s specialization in rare physiologically complex surgical procedures. Anesthesiology 2005;103:161–7.[ISI][Medline]
  5. Dexter F, O’Neill L. Data envelopment analysis to determine by how much hospitals can increase elective inpatient surgical workload for each specialty. Anesth Analg 2004;99:1492–500.[Abstract/Free Full Text]
  6. O’Neill L, Dexter F. Market capture of inpatient perioperative services using data envelopment analysis. Health Care Manag Sci 2004;7:263–73.[Medline]
  7. O’Neill L, Dexter F. Methods for understanding super-efficient data envelopment analysis results with an application to hospital inpatient surgery. Health Care Manag Sci 2005;8:291–8.[Medline]
  8. Dexter F, O’Neill L. Tactical increases in operating room block time based on financial data and market growth estimates from data envelopment analysis. Anesth Analg 2007;104:355–68.[Abstract/Free Full Text]
  9. Macario A, Dexter F, Traub RD. Hospital profitability per hour of operating room time can vary among surgeons. Anesth Analg 2001;93:669–75.[Abstract/Free Full Text]
  10. Dexter F, Blake JT, Penning DH, Lubarsky DA. Calculating a potential increase in hospital margin for elective surgery by changing operating room time allocations or increasing nursing staffing to permit completion of more cases: a case study. Anesth Analg 2002;94:138–42.[Abstract/Free Full Text]
  11. Dexter F, Ledolter H, Wachtel RE. Tactical decision making for selective expansion of operating room resources incorporating financial criteria and uncertainty in sub-specialties’ future workloads. Anesth Analg 2005;100:1425–32.[Abstract/Free Full Text]
  12. Luft HS, Garnick DW, Mark DH, et al. Does quality influence choice of hospital. JAMA 1990;263:2899–906.[Abstract]
  13. Finlayson SR, Birkmeyer JD, Tosteson AN, et al. Patient preferences for location of care: implications for regionalization. Med Care 1999;37:204–9.[ISI][Medline]
  14. Iowa Hospital Association, Available at www.ihaonline.org/publications/trustee%20resources/iha.pdf. Accessed September 2006.
  15. Hospitals in New York State—-Regional Offices, Available at www.nyhealth.gov/nysdoh/hospital/regional_offices.htm, and New York State Hospital Profile, http://hospitals.nyhealth.gov/. Accessed September 2006.
  16. Adapted from the Official New York State Tourism Website, Available at www.iloveny.com/info_center/pdf/map_nys_base.pdf. Accessed September 2006.
  17. New York State Department of Health Statewide Planning and Research Cooperative System Health Service Areas, Available at www.health.state.ny.us/statistics/sparcs/sysdoc/appu.htm. Accessed September 2006.
  18. New York State Department of Transportation Regional Offices, Available at www.dot.state.ny.us/reg/regmenu.html. Accessed September 2006.
  19. New York - Visitors Network New York State Regions, Available at www.visitnewyorkstate.net/regions/. Accessed September 2006.
  20. Hahn GJ, Meeker WQ. Statistical intervals. A guide for practitioners New York: Wiley, 1991: 123.
  21. New York State Department of Health Statewide Planning and Research Cooperative System, Available at www.health.state.ny.us/statistics/sparcs/index.htm. Accessed September 2006.
  22. Horan TC, Emori TG. Definitions of key terms used in the NNIS system. Am J Infect Control 1997;25:112–6.[ISI][Medline]
  23. American Society of Anesthesiologists’ Relative value guide. Park Ridge, Illinois: American Society of Anesthesiologists, 1999.
  24. Dexter F, Thompson E. Relative value guide basic units in operating room scheduling to ensure compliance with anesthesia group policies for surgical procedures performed at each anesthetizing location. AANA J 2001;69:120–3.[Medline]
  25. Dexter F, Macario A, Penning DH, Chung P. Development of an appropriate list of surgical procedures of a specified maximum anesthetic complexity to be performed at a new ambulatory surgery facility. Anesth Analg 2002;95:78–82.[Abstract/Free Full Text]
  26. Clinical Classifications Software (CCS) for ICD-9-CM. Available at www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp. Accessed October 2004.
  27. Kanter RK, Dexter F. Criteria for identification of comprehensive pediatric hospitals and referral regions. J Pediatr 2005; 146:26–9.[ISI][Medline]
  28. Map of University Hospital’s 13-County Primary, Secondary and Tertiary Service Areas. Available at www.upstate.edu/mcbd/mark_res/maps.shtml. Accessed September 2006.
  29. United States Census Bureau Population Estimates. Available at www.census.gov/popest/datasets.html. Accessed September 2006.
  30. United States Census Bureau Small Area Income and Poverty Estimates. Available at www.census.gov/hhes/www/saipe/saipe.html. Accessed September 2006.
  31. United States Census Bureau Small Area Health Insurance Estimates. Available at www.census.gov/hhes/www/sahie/data.html. Accessed September 2006.
  32. Iowa Hospital Association Hospital and Health System Characteristics. Available at www.ihaonline.org/publications/profileserv/I%20Hospital%20and%20Health%20System%20Characteristics.pdf. Accessed September 2006.
  33. United States Census Bureau County-to-County Worker Flow Files. Available at www.census.gov/population/www/cen2000/commuting.html. Accessed September 2006.
  34. Tulloh B, Clifforth S, Miller I. Caseload in rural general surgical practice and implications for training. ANZ J Surg 2001;71: 215–7.[Medline]
  35. Gates RL, Walker JT, Denning DA. Workforce patterns of rural surgeons in West Virginia. Am Surg 2003;69:367–71.[ISI][Medline]
  36. Hughes-Anderson W, House J, Aitken RJ, et al. Analysis of the outcomes of a visiting surgical service to small rural communities. ANZ J Surg 2003;73:833–5.[ISI][Medline]
  37. Sheikh K. Utility of provider volume as an indicator of medical care quality and for policy decisions. Am J Med 2001;111:712–5.[ISI][Medline]
  38. Papadimos TJ, Habib RH, Zacharias A, et al. Early efficacy of CABG care delivery in a low procedure-volume community hospital: operative and midterm results. BMC Surg 2005;5: 10–8.[Medline]
  39. American Hospital Association, Small or Rural Hospitals. Available at www.hospitalconnect.com/aha/member_relations/cah/hospital_assistance.html. Accessed September 2006.
  40. Dimick JB, Finlayson SR, Birkmeyer JD. Regional availability of high-volume hospitals for major surgery. Health Aff Suppl Web Exclusive 2004;VAR45–VAR53.
  41. Bristow RE, Zahurak ML, del Carmen MG, et al. Ovarian cancer surgery in Maryland: volume-based access to care. Gynecol Oncol 2004;93:353–60.[ISI][Medline]
  42. Birkmeyer JD, Siewers AE, Marth NJ, Goodman DC. Regionalization of high-risk surgery and implications for patient travel times. JAMA 2003;290:2703–8.[Abstract/Free Full Text]
  43. Procedures in U.S. Hospitals, 1997, HCUP Fact Book No. 2. Available at www.ahrq.gov/data/hcup/factbk2/factbk2.pdf. Accessed September 2006.
  44. Baker LC. Measuring competition in health care markets. Health Serv Res 2001;36:223–51.[ISI][Medline]



This article has been cited by other articles:


Home page
 Anesth. Analg.Home page
F. Dexter and R. H. Epstein
Calculating Institutional Support That Benefits Both the Anesthesia Group and Hospital
Anesth. Analg., February 1, 2008; 106(2): 544 - 553.
[Abstract] [Full Text] [PDF]

Home page
 Anesth. Analg.Home page
T. McCall
Green Mountain Boys Would Be Disappointed
Anesth. Analg., October 1, 2007; 105(4): 1169 - 1170.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a colleague
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles