PEDIATRICS Vol. 110 No. 4 October 2002, pp. 849-850
Can Regionalization Decrease the Number of Deaths for Children Who Undergo Cardiac Surgery? A Theoretical Analysis
To the Editor.—Chang and Klitzner1 theorize that "regionalization" of services may result in decreased mortality in children undergoing cardiac surgery. This study raises several questions and provokes commentary.
The authors suggest their analysis may not be applicable to other areas of the United States. Consider the implications if it were. Applying the model to a national population of roughly 281 million2 would "avoid" the deaths of over 400 children each year. The scenario proposed by the authors would regionalize care so that a reduced number of high-volume centers would service an average of 6.8 million people and perform roughly 400 surgical procedures. Such a model would, however, mean that 75% of the states would not have a population sufficient to support a surgical program. We ask the authors to comment on the following. What options are available to a "spelling" 2-month-old infant with tetralogy of Fallot who lives in South Dakota or New Mexico? What would be the effect of national regionalization on travel distance? What agency would oversee implementation and regulation of this regionalization process? What reaction could be anticipated from state legislators, public health officials, hospital administrators, medical school deans, department chairmen, etc, whose programs would be terminated? Where would the nurses, respiratory therapists, child-life specialists, echo-technicians, surgeons, and cardiologists, etc, find new employment?
If we accept the authors’ hypothesis that volume alone stands as the variable influencing mortality, we can then reasonably apply their theoretical construct in the opposite direction. Shifting patients from high-volume programs to low-volume programs would decrease the variance about the mean and improve the observed mortality. Applying this scenario to the authors’ data, we propose closing the low-volume, poor-outcome programs (mortality >0.06) and the 3 ultra-high-volume programs (>250 case per year). Surgical cases would be shifted to yield 11 programs performing roughly 190 cases each year. The number of avoidable deaths would be statistically equivalent to that proposed by the authors. Travel distance would be lessened, and patient choice would be enhanced.
This scenario would also limit the capacity of third-party payers to "commoditize" clinical service and diminish downward pricing pressure upon programs with the threat of movement of a large number of plan subscribers from one center to another. Clinicians and administrators quick to embrace the concept of regionalization should be reminded of the gallows humor circulated within so-called "centers of excellence": "With this contract we’ll lose $2000 per case—but we’ll make it up on volume."
Few would argue the absurdity of our proposed scenario. As representatives of an intermediate-volume, high-quality program, we offer this proposal simply to illustrate the blatant self-serving nature of the regionalization concept. The authors, as clinicians working at a high-volume program that would directly benefit from the implementation of such a proposal, should acknowledge the same.
Volume alone is not the sine qua non of outcome. Successful pediatric/congenital cardiovascular programs evolve with the cooperation of dedicated, skilled, intelligent, persistent people. Through such efforts, quality outcomes are achieved, and program volume grows. In short, quality creates volume, not the other way around. Research efforts are needed to identify those systems, operations, and protocols that produce high quality so the resultant knowledge and insight can be applied for the benefit of all programs, patients, and families.
Philip C. Smith, MD, PhD
The Heart Center
Division of Cardiovascular Surgery
Children’s Hospital Medical Center of Akron
Akron, OH 44308
Keith R. Powell, MD
Department of Pediatrics
Children’s Hospital Medical Center of Akron
Akron, OH 44308
REFERENCES
- Chang R-KR, Klitzner TS. Can regionalization decrease the number of deaths for children who undergo cardiac surgery? A theoretical analysis.
Pediatrics.2002; 109
:173
–181
[Abstract/Free Full Text] - USA Statistics in Brief—Census 2000 Resident Population of States. US Census Bureau. December 28, 2000
In Reply.—
We appreciate the interest in our paper displayed by Smith and Powell. Their questions reflect frequently raised concerns regarding regionalization. We hope to further clarify some important issues.
As pediatric cardiac specialists, our ultimate goal is to provide quality care and improve outcomes for children with heart disease. As our study demonstrates,1 regionalization of services by referring patients from low-volume hospitals to high-volume hospitals is one way to achieve these goals. However, there are many other ways in which quality of care for children with heart disease can be improved. In the discussion section of our article, we suggest that quality assurance programs, selective use of clinical pathways, and continuing education of medical personnel can all contribute to improvements in quality of care. In evaluating these strategies, both cost-effectiveness and access to care must be considered. Elucidation of these complex issues is beyond the scope of this dialogue and requires additional investigation.
We agree with the implied caveat of Smith and Powell that extrapolation of our study’s results to geographic areas not similar to California requires care and thought. Although California contains both urban and rural areas, the distance from the most rural areas to the nearest urban center in the state is shorter than would be found in some very sparsely populated regions of the country. In addition, preferred modes of transportation, traffic conditions, and travel habits of local residents vary across the country. The definition of high- and low-volume hospitals may also differ by region. Many of us would agree that healthcare resources are not evenly distributed. Instead of having many pediatric cardiac centers, as is the case in California, many states may have only 1 pediatric cardiac center, or none at all. These states face special challenges with respect to regionalization. Nonetheless, the interpretation of our results, that selective referral to centers performing >400 cases per year is the best mechanism to achieve better outcomes in all regions, is not only tainted with impracticality but also misleading. Regionalization is a process that should be conducted on a region-by-region basis, taking into account the characteristics of local geography, demographics, and healthcare markets.
From the perspective of healthcare policy, we believe that regionalization as a means to improve quality of care should not compromise accessibility of services and the availability of patient choice. Although in our study there was no dramatic increase in travel distance until most of the hospitals in California were "closed," this may not be the case for other areas of the United States. In some states or regions, combining programs may increase patient travel distance by hundreds of miles.
Regionalization can be accomplished by a hierarchical ordering of services. Although eliminating low-volume programs may not be practical, we have included an analysis of what may be a more feasible solution, namely selective referral of "high-risk" patients to high-volume centers. In our analysis, this would require transfer of 11% of patients, or 27% of patients in low- and medium-volume hospitals, to high-volume hospitals. This strategy will potentially reduce the mortality rate to 4.6% without a dramatic increase in mean travel distance. In addition, accessibility of services in an emergent situation will not be compromised. For example, a patient with tetralogy of Fallot should have a repair (high-risk procedure) at a high-volume hospital. In contrast, a "spelling" 2-month-old with tetralogy of Fallot may face lower risk by having an aortopulmonary shunt (low-risk procedure) performed at the nearest center regardless of its high or low volume. From our results, problems will occur when low-volume centers consistently perform high-risk operations, thereby driving up the overall mortality rate for a given region.
The relation between case volume and surgical outcomes in pediatric cardiac surgery has been well documented.2,3 However, the explanation for this observation remains controversial. The basis of debate hinges on the question: Do high case volumes lead to better outcomes ("practice makes perfect"), or do superior outcomes attract more referrals, resulting in higher case volume. Our study does not require the validity of either theory. We utilized the volume-outcome relation as the basis for referral regardless of whether high volume is the cause or consequence of better outcomes. We submit that our results are valid independent of the cause of the volume-outcome relationship.
We acknowledge that case volume alone may not be the only indicator for the quality of care in a given hospital. However, higher case volume is associated with hospital characteristics known to result in a better quality of care, such as availability of adequately staffed and equipped operating rooms, intensive care units, cath labs, ancillary services, new technologies, and training programs.4 We believe that most pediatric cardiac specialists would concur there is a need to develop better systems to assess the quality of care for pediatric cardiac services. Until these specific measures of quality are available, case volume will remain the only proven indicator which is available and easily quantifiable.
Finally, we wish to stress that volume and outcome relation is a statistical association yielding the impression that high-volume hospitals generally have better outcomes. As is true for any statistical analysis, there are always outliers. As shown in the scatterplot in Fig 2 of our paper, there are some low- and medium-volume hospitals that have low mortality rates, some even lower than some high-volume centers. We agree with Smith and Powell that many high-quality programs exist independent of their surgical volume. We feel strongly that future research efforts should aim to examine other variables potentially related to outcomes. We hope to engage in such studies in the future to identify the characteristics of high-quality, low-volume hospitals that distinguish them from other low-volume hospitals with high mortality rates.
Ruey-Kang R. Chang, MD, MPH
Division of Cardiology, Department of Pediatrics
Harbor-UCLA Medical Center
Torrance, CA
Thomas S. Klitzner, MD, PhD
Division of Cardiology, Department of Pediatrics
UCLA School of Medicine
Los Angeles, CA
REFERENCES
- Chang R-KR, Klitzner TS. Can regionalization decrease the number of deaths for children undergoing cardiac surgery? A theoretical analysis. Pediatrics.2002; 109 :173 –181
- Jenkins KJ, Newburger JW, Lock JE, et al. In-hospital mortality for surgical repair of congenital heart defects: preliminary observations of variation by hospital caseload.
Pediatrics.1995; 95
:323
–330
[Abstract/Free Full Text] - Hannan EL, Racz M, Kavey RE, Quaegebeur JM, Williams R. Pediatric cardiac surgery: the effect of hospital and surgeon volume on in-hospital mortality.
Pediatrics.1998; 101
:963
–969
[Abstract/Free Full Text] - American Academy of Pediatrics Section on Cardiology and Cardiac Surgery. Guidelines for pediatric cardiovascular centers.
Pediatrics.2002; 109
:544
–549
[Abstract/Free Full Text]
PEDIATRICS (ISSN 1098-4275). ©2002 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
