| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Division of Cardiothoracic Surgery, University of Washington Medical School, Seattle, Washington
b Department of Surgery, University of Washington Medical School, Seattle, Washington
c Surgical Outcomes Research Center, University of Washington Medical School, Seattle, Washington
Accepted for publication November 8, 2010.
* Address correspondence to Dr Varghese, Section of Thoracic Surgery, University of Washington, 1959 NE Pacific, Suite AA-115, Box 356310, Seattle, WA 98195 (Email: tkv{at}u.washington.edu).
Presented at the Fifty-sixth Annual Meeting of the Southern Thoracic Surgical Association, Marco Island, FL, Nov 4–7, 2009.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Methods: Using the Washington State Comprehensive Hospital Abstract Reporting System, a retrospective cohort design evaluated all patients (
18 years) undergoing esophageal resection for any diagnosis since the introduction of Leapfrog standards (2000 to 2007). The main outcome measures were hospital stay, readmissions within 30 days of discharge, discharge to an institutional care facility, operative reinterventions, and 90-day mortality.
Results: A total of 1,505 adult Washington state residents underwent esophageal resection without complex reconstruction (1,352 elective [89.8%]). Of 45 hospitals reporting at least one procedure, 5 (11%) met Leapfrog volume standards. Leapfrog hospitals accounted for 62% of the total elective volume. Overall, elective patients at Leapfrog hospitals had a lower adjusted risk of death compared with those at hospitals that did not meet criteria (odds ratio 0.50, p = 0.02). Across the different Leapfrog hospitals there was over fivefold variation in 90-day mortality (1.7% to 10.2%), 2.5-fold variation in reinterventions (8% to 20%), and fourfold variation in discharges to an institutional care facility (5.3% to 19.8%). Length of stay and readmission rate varied less.
Conclusions: Although referral to high-volume centers has been an important advance for complex surgical procedures, there is still a substantial degree of variability in outcomes among hospitals that met Leapfrog volume criteria for esophagectomy. Metrics such as process, individual surgeon volume, and risk-adjusted outcome measures may yield further opportunities for quality improvement that extend beyond hospital volume-based assessments.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
| GENERAL THORACIC SURGERY:
The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
|
Several studies have demonstrated a significant relationship between volume and outcome for specific surgical procedures. In response, payers, policy makers, and professional organizations have implemented a variety of strategies aimed at improving the quality of surgical care [1]. One such strategy is surgical volume-based referral for complex procedures. Esophageal resections have been identified as an example of a complex procedure where a volume outcome relationship exists. Despite advances in perioperative care and surgical techniques, the mortality rate after esophagectomy remains unacceptably high, with a worldwide rate of 11% [2]. On a nationwide basis, esophagectomy carries considerable, but variable, operative risk. A landmark study in 2002 cited operative mortality varying from 8% at "high-volume" centers to 23% at "low-volume" centers [3]. A more recent study cited overall mortality rate for esophagectomy in the US to be 7% in 2006 [4].
In response to an Institute of Medicine report on building a safer health system [5], several large employers formed the Leapfrog Group (LF) in November 2000. The objective of the group was to improve the quality and safety of medical care. Leapfrog identified three potential "leaps" forward in the area of patient safety that were directed specifically at hospitals: "computer physician order entry" for medications in hospitals, the use of intensivists in hospital intensive care units, and referring patients to hospitals that meet volume (or outcomes where available) thresholds and process standards [6]. They chose these particular leaps because they believed they were supported by research, their adoption could make a measurable improvement in safety, and because the leaps had intuitive appeal to the general public. Some have felt that the Leapfrog Group, to date, is the most ambitious coordinated attempt on the part of large employers to reshape healthcare in America [7]. The resulting initiative thus suggested selective referral of complex procedures to high-volume hospitals and set volume thresholds for five procedures. The Leapfrog Group standards for evidence based referral underwent significant revision in 2003. The Leapfrog 2003 standards were based on minimum volume standards alone for two operations (esophagectomy and pancreatectomy) and volume standards plus other patient outcomes or processes of care for three other procedures (coronary artery bypass grafting, abdominal aortic aneurysm repair, and percutaneous coronary intervention) [8]. A high-volume center for esophagectomy is defined by the Leapfrog Group as one that performs greater than 13 per year. The purpose of this study was to determine whether meeting the case-volume benchmark was consistently associated with improved patient outcomes for esophagectomy in the state of Washington.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Setting and Data Source
The Washington State Comprehensive Hospital Abstract Reporting System (CHARS) database contains records of all hospitalizations occurring between 2000 and 2007, except those at Veterans Affairs and US military hospitals. This data set includes the following: demographic information (age and sex); International Classification of Diseases, Ninth Revision (ICD-9) diagnostic and procedure codes; administrative details of hospitalization (insurance status, length of stay, charges, and disposition); and coded hospital identifiers. Ninety-day mortality was determined by means of linkage of the CHARS to the Washington State Department of Health vital records database.
Patient Sample
Patients aged 18 and older who underwent esophageal resection in the state of Washington between 2000 and 2007 were identified from the CHARS database (Fig 1
). The following ICD-9 codes were used to identify patients undergoing resection: Esophagectomy (42.40–42.42); esophagogastrectomy not otherwise specified (43.99); intrathoracic esophagogastrostomy (42.52); antesternal esophagogastrostomy (42.62); and partial gastrectomy with anastomosis to esophagus (resection of the gastroesophageal junction) (43.5). Excluded from analysis were patients who had out of state zip codes, and hence potential for inaccuracy of follow-up data, and those patients who underwent complex reconstruction (ICD-9 codes 42.43, 42.55, 42.58, 42.63, 42.65, 42.68, 43.91). Patients with a diagnosis of esophageal cancer were identified by ICD-9 codes 150.0 to 150.9 and 151.0. The analysis population consisted of 1,352 patients who underwent elective esophageal resections.
|
13 vs 1 to 12). This was determined using the annual average number of esophageal resections conducted during 2000 to 2007. Leapfrog hospitals were ranked by esophageal resection volume.
Outcomes
Outcomes of interest included length of stay, prolonged length of stay (PLOS; greater than 14 days), readmission within 30 days after initial discharge after surgery, discharge to an institutional care facility (ICF), reintervention, and 90-day mortality. Only those who survived greater than 14 days were included in the analysis for PLOS, while readmission rates and discharge to ICF were determined among in-patient survivors. Discharge to an ICF indicated a patient who went to another facility such as a skilled nursing home, intermediate-care facility, or a rehabilitation facility versus home. Ninety-day mortality was defined as "all-cause" death within 90 days of discharge as ascertained from Washington State Vital Records. Reintervention was defined as those procedures performed for complications after esophageal resection, and included the following ICD-9 codes: Tracheostomy (31.1 to 31.2X); other incision of larynx or trachea (313); reopening of thoracotomy site (34.03); reoperation for bleeding (39.41, 39.49, 39.98); ligation of thoracic duct (40.64); reopening of operative site (54.12); laparotomy for drainage of intestinal or peritoneal abscess (54.19); and percutaneous abdominal drainage (54.91).
Covariates
The Deyo modification of the Charlson comorbidity index was calculated by using ICD-9 diagnostic codes from index admissions [9]. A value of zero indicates no comorbid conditions, whereas an index of 3 or greater indicates the greatest extent of comorbidity. Insurance status was considered a marker for economic status and was a dichotomized variable indicating whether the primary payer was Medicaid or the case was considered "charity" versus all other types of payers, such as Medicare, a health maintenance organization, Blue Cross-Blue Shield, or commercial insurance.
Analysis
The Stata program (special edition 9.2; Statacorp, College Station, TX) was used for all statistical analyses. The Pearson 
2 test was used to compare categoric variables and a nonparametric equality of medians test was used to compare continuous variables. Confidence intervals for binary outcomes were obtained using exact binomial methods. Logistic regression models were used to examine the relationship between hospital Leapfrog status and binary outcomes. The unit of analysis was the patient and all models adjusted for clustering at the hospital level.
Models were adjusted for age, sex, Charlson comorbidity index, indication for resection (benign versus malignant), insurance status, and calendar year. One patient (0.07% of the cohort) had a missing value for insurance and was excluded from adjusted analyses.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
|
|
|
Unadjusted short-term outcomes are shown in Table 3. The top two volume hospitals had shorter lengths of stay as compared to non-Leapfrog hospitals, while the others had equal or greater lengths of stay. Adjusted short-term outcomes are shown in Tables 4 and 5. Leapfrog hospitals overall had 50% lower odds of mortality within 90 days (Table 4), consistent with other studies in the literature. Although there were trends toward lower odds of PLOS, reinterventions, and discharge to ICF, these did not meet statistical significance. Surprisingly we found 28% higher odds of readmissions for those patients who underwent esophageal resections at Leapfrog hospitals.
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
Questions have been raised about the ramifications of policy making based on case volume. Although there appears to be a statistical link between volume and quality of care, the nature of this link is poorly understood. A recent retrospective single-center study [11] compared morbidity and mortality at their low-volume center with regional high-volume centers for esophagectomy and suggested that volume alone is not a sufficient signal of quality. Additional factors that can influence outcomes include surgeons' skill and experience, and the presence of an organizational structure for assuring high quality of care such as organized multidisciplinary management and treatment protocols. It is possible that for the most complex procedures, whether frequent or infrequent, hospitals providing them must maintain a certain level of staffing and technology to support even one procedure. In addition, a few specialized surgeons might perform these procedures at more than one hospital, so an individual hospital's volume, whether high or low, may be a poor proxy for outcomes. Individual surgeon training or volume, hospital staffing, or measures of the presence of key technologies or practices, such as protocols, may be better measures. There is also a complex interplay of forces. For example, high-volume surgeons at low-volume centers perform better as do low-volume surgeons at high-volume centers [12].
There has been a recent significant interest in using indicators of increased healthcare utilization as a surrogate for quality under the assumption that adverse events occur more frequently among patients whose hospitalizations are marked by greater use of resources. One such surrogate measure, PLOS greater than 14 days, has been linked to significantly more complications and higher mortality [13]. The National Quality Forum endorsed risk-adjusted PLOS as a hospital-level quality metric on August 8, 2008 [14]. However, any one indicator of healthcare use may miss the opportunity to describe overall resource utilization. In addition to PLOS, our study examined the use of two other indicators of increased healthcare utilization; discharge to ICF and readmission rates. We found no difference in discharge to ICF and were surprised by the finding of increased odds for readmission in Leapfrog hospitals. Reasons may include higher distances travelled by patients to go to a Leapfrog hospital as often readmissions can occur at a hospital other than one where the resection was performed [15]. There is the possibility that there is a low threshold to admit patients for fairly routine postoperative problems if they present to an outside hospital, without direct knowledge of the individual patient or experience in postoperative management of the surgery itself. This is a potential unintended consequence of regionalization of high-risk procedures.
There is also the question of whether the higher readmission rate in Leapfrog compliant hospitals is due to an improved recognition of postoperative complications and their management. A recent study demonstrated that an additional factor in decreased mortality rates after in-patient surgery is how well hospitals recognize and rescue patients from major complications once they have occurred [16]. If there are no differences in clinically important complications after surgical procedures between high-volume and low-volume hospitals, is increased healthcare utilization the price to pay for rescuing these patients? These areas of uncertainty warrant further investigation, ideally with a prospectively collected, reliable, and valid data source with the ability to construct a risk-adjusted model that accounts for clinical and nonclinical determinants of healthcare utilization. This retrospective study is limited in that we were not able to discern the exact reason for readmission and whether the hospital to which the patient was readmitted to was different from the original hospital where the resection occurred.
In our study, although the Leapfrog hospitals overall had improved outcomes compared with non-Leapfrog hospitals, many low-volume centers had acceptable performance while some high-volume hospitals had performances no different than other low-volume poor performers. This variability calls into question the value of the current Leapfrog volume criteria in helping patients and insurers make decisions about where to get high-risk operations, and may end up regionalizing care away from centers with adequate results to centers that perform a high volume of cases with worse results. However, the top performing hospitals were the two highest volume Leapfrog compliant centers and had substantially better outcomes as measured by mortality, PLOS, reintervention rates, discharge to ICF, and median hospital LOS. Thus the current Leapfrog volume criteria may be in fact too low. A recent national study that examined the independent effect of annual case volume on complication rates, demonstrated that improved mortality rates occurred with higher volume until it leveled off at 30 esophagectomies per year [4]. Using these criteria, only the two highest volume centers in the state of Washington would qualify as ideal centers. Clearly if patients needing esophagectomy could be cared for in these high-performing hospitals, or if the processes of care that result in improved outcomes could be delivered effectively in the other hospitals performing esophagectomy, outcomes for patients would improve dramatically.
In addition to structural measures such as volume, process-of-care or direct outcome measures could be used to signify high-quality hospitals. Process analysis may be a more productive approach to quality improvement projects than outcomes-oriented approaches. Some procedures may achieve superior outcomes with specific processes of care. Identification of these processes, and exporting them in educational efforts, can then lead to overall improvement in results. The problem is that there are a limited number of evidence-based processes of care measures that link to outcome and virtually none of these are well established for esophageal surgery.
This study has several limitations. The CHARS database did not include patients in military hospitals, Veterans Administration medical centers, or Washington residents who underwent esophageal resection in the state and immediately moved to a different state. Administrative data sets do not contain sufficient clinical information to adequately adjust for potential confounding based on the stage of cancer or underlying severity of comorbid conditions. The goal of our study was to describe variation in outcome rather than predict future performance. Chance is an important consideration when describing variation in outcomes across centers. The role chance may be playing in our findings is reflected by the wide confidence intervals associated with the estimates. However, for esophageal resections, Leapfrog uses only volume metrics to label an institution as a high-volume or low-volume center, and adjustments based on past performance is not part of the Leapfrog criteria. Explaining the persistent variation and development of alternative predictive measures, while essential to a quality improvement project, are beyond the scope of this descriptive study.
The CHARS database also does not have esophageal cancer specific information with regards to staging, pathology, or neoadjuvant therapy. However, we felt that though these factors may impact long-term survival in the treatment of esophageal cancer, their influence on the immediate perioperative outcomes examined in the present study would be limited.
In summary, a high-quality hospital as defined by the current Leapfrog definition ( 13 esophagectomies per year) appears to be inconsistent. Volume-based referrals alone are counting on what appears to be an indirect and complex link between increased case volume to better outcomes. This may reflect the organization of healthcare services, patient selection, and the skill and experience of staff. Solely relying on hospital volume numbers to determine quality of esophagectomy care can lead to a false sense of security that best outcomes can be assured with volume-based referral. Efforts to improve surgical quality will need to look beyond the current volume-based referral alone.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
This study highlights the important and elusive concept of how to optimize our ability to monitor and improve the quality of care in thoracic surgery. Of all the procedures we perform in general thoracic surgery, esophagectomy has been demonstrated to be the most sensitive to surgeon training, surgeon volume, and hospital volume. The focus of your analysis is a standard proposed by the Leapfrog group, a standard not met by 89% of the hospitals in the state of Washington, affecting approximately 40% of the esophagectomy patients in your state, a standard apparently refuted by your study, although I would point out that the Leapfrog standards aren't meant to guarantee better outcomes; and I will point out a recent meta-analysis of over 32 studies published by Gruen and colleagues in Cancer Journal for Clinicians this month, and of those 32 studies, more than three-quarters demonstrated surgeon and (or) both hospital volume as being linked to outcome. So, in the absence of risk stratification, which I think might clarify some of the issues in this study, in the absence of that risk stratification it is difficult to make any conclusions regarding the outcomes of esophagectomy in your state, but you have the power to do that more locally. So my first question is, how many surgeons perform esophagectomy in your group and what is the range of volumes among these surgeons? How do you manage practice and outcome variations in your own department?
Number two, you appear not to agree with the Leapfrog standard of 13, which in a medium-sized operative group might mean that surgeons do three or less esophagectomies a year. So are there other standards other than volume that you would recommend today to improve outcomes?
And lastly, in the Gruen et al meta-analysis that I just referred to, it was stated that of all gastrointestinal operations esophagectomy is the most linked to volume and that it is estimated that it would only take 10 patients to be transferred to a higher volume center to save one life, which is a staggeringly low number of patients. So, in light of that regarding this Leapfrog standard and the results of your study, do you believe that esophagectomy should be regionalized and what other information or analysis would be required to establish that policy decision either in your state or in the country?
Thank you.
DR VARGHESE: Thanks for the questions, Dr D'Amico. Within our own university we have both thoracic surgeons and general surgeons performing esophageal resections. The current study is an example of collaboration between both groups. Currently, the outcomes for thoracic surgeons are captured in the STS [Society of Thoracic Surgeons] database but the outcomes of the general surgeons are not. There is variation in practices between both groups and we are in the process of prospectively analyzing the outcomes for the groups. We also wish to expand this prospective analysis to the state of Washington. There is a surgical outcomes program called SCOAP in the State of Washington; the Surgical Clinical Outcomes Assessment Program. This initially started with just general surgical procedures, examining whether process measures such as timely administration of antibiotics, receipt of blood transfusion, and normothermia during the case would affect outcomes. As of that project, 56 hospitals were recruited in this grass-roots quality initiative in the State of Washington. What we are hoping to do is use the SCOAP program as a template for quality improvement. An essential first step will be for consensus agreement on the metric process measures.
With regards to your question of the relationship between volume and outcomes, we don't completely disagree with Leapfrog. I think Leapfrog was an important first step. The only problem that we have is, unlike all the other procedures that were listed out in terms of Leapfrog standards, for esophagectomies and pancreatic resections volume alone is the only determinant of quality. There are bound to be process measures that can be used to drive quality for esophageal resections. What we need is to determine what those measures are.
In terms of regionalization, the meta-analysis you cited stated that moving 10 patients to a high-volume center could save one life. I think in the ideal setting we do agree that regionalization makes sense. The problem is from the practical aspect. In a number of these studies where patterns of patients being referred to high-volume centers have been analyzed, it is the patients that are left behind, such as the poor or the minorities, who get severely disadvantaged. In a setting where they are already having difficulty for access to care, do you end up under-serving that population even more? Hence, in addition to regionalization, it may also be important to try and improve quality all across the board. That is, to make sure that if any surgeon performs an esophageal resection there should be a baseline level of expertise, as well as collection of key process measures in order to look at outcomes in a constructive manner.
DR WILLIAM A. COOK (North Andover, MA): I would like to address the issue of regionalization from the standpoint of my practice over the last 35 years in the Merrimack Valley. When I first went there, I did virtually everything I was doing at the Albert Einstein in New York, and over time, more and more of that material has been going to Boston for a number of reasons, not to get too detailed here. The net effect of that is not just on my practice but it is on the practice of everyone who has a community type practice. So, a year and a half ago a young man came to practice in the Merrimack Valley and after a year he left because he simply couldn't find enough work to make a viable practice. So that led to further under-serving of thoracic needs in the Merrimack Valley. But it has a second effect, and I think this is more important for us all to hear.
What has happened in Massachusetts is that, whereas the national average of medical care in tertiary institutions is 16%, in Massachusetts it is 40%. In other words, 40% of all the medical care that is being delivered in Massachusetts is being delivered in the tertiary hospitals. And as you can imagine, it costs about three times as much to do a procedure, any thoracic procedure that I might do, in Boston over what it would cost to do it at home, and it has resulted in Massachusetts having the highest cost of medical insurance in the country. It doesn't take a genius to figure out how that would be.
So, you know, over the years I have been trying to think, well, is there a way to change that? And I am wondering if anybody has ever thought about the business of taking your expertise and going out to me and operating with me in my hospital, which is perfectly well set up to do the things that I do or I wouldn't do them, so that there is a dialogue, number one, between the teaching institutions and the bigger hospitals and the people in the smaller communities, which would serve the purpose of driving that economic thing in the opposite direction, number one, and, number two, would absolutely help in the development of the community surgeon, new things like, let's say, doing VATS [video-assisted thoracic surgery] type lobectomies and things like that, which one might not want to start to do alone in a community hospital. So you would serve two very important purposes towards the total education of all the thoracic surgeons in America and not just the guys who are going to be in those "regional centers." I think there is something to think about in that.
DR VARGHESE: Thanks, Dr Cook. You do make a good point. Obviously a major limitation of regionalization is cost. That is, does it cost more to move patients to a high-volume center, and do you counterbalance that by the savings in patient safety (better outcomes). One explanation for the higher readmission rate in our study could be the concept of "failure to rescue" that was recently published in the New England Journal from the group in Michigan. In this theory the overall complication rates at a high-volume center and a low-volume center may indeed be exactly the same, and that the difference you are seeing is that in the low-volume center they "fail to rescue" that patient from that complication leading to higher mortality. The trade-off for saving that patient may indeed be a higher readmission rate. Current studies are underway to analyze this concept. With the recent emphasis on healthcare reform, treatment comparative effectiveness and cost analyses are additional issues we will need to analyze in the future.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. R. Gopaldas, C. M. Bhamidipati, T. K. Dao, and J. G. Markley Impact of Surgeon Demographics and Technique on Outcomes After Esophageal Resections: A Nationwide Study Ann. Thorac. Surg., March 1, 2013; 95(3): 1064 - 1069. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. T. Cooke and D. H. Wisner Who Performs Complex Noncardiac Thoracic Surgery in United States Academic Medical Centers? Ann. Thorac. Surg., October 1, 2012; 94(4): 1060 - 1064. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Harrison Assessing the relationship between volume and outcome in hospital services: implications for service centralization Health Services Management Research, February 1, 2012; 25(1): 1 - 6. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Harrison Assessing the relationship between volume and outcome in hospital services: implications for service centralization Health Serv Manage Res, February 1, 2012; 25(1): 1 - 6. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
