The Relationship between Hospital Case Volume and Outcome from Carotid Endartectomy in England from 2000 to 2005

Article Outline

• Abstract
• Introduction
• Methods
  • Data quality
• Statistical Analysis
  • Mortality-control charts
  • Volume analysis
  • Multi-factorial analyses
  • Hospital safety assessment
• Results
  • Summary data
  • Mortality control charts
  • The effect of volume
  • Hospital numbers
  • Complication rate and length of stay
  • Risk of uni-/bi-lateral stenoses
  • Safety charts
  • Age and gender
  • Year of procedure
• Discussion
  • Case-mix
  • Stroke
• Conclusions
• References
• Copyright

Objectives

To assess the outcome of carotid endarterectomy in England with respect to the hospital case-volume.

Methods

Data were from English Hospital Episode Statistics (2000–2005). Admissions were classified as elective or emergency. Risk-adjusted data were analysed through modelling of death rate, complication rate and length of admission with regard to the year of procedure and annual hospital volume of surgery. Hospitals with elevated death rates were identified and the evidence quantified that they had outlying mortality rates.

Results

There were 280 081 diagnoses of extra-cranial atherosclerotic arterial disease in which 18 248 CEA were performed. The mean mortality rates were 1.04% for elective and 3.16% for emergency CEA.

A volume-related improvement in mortality (p=0.047) was seen for elective CEA. Length of stay decreased as annual volume increased for elective and emergency CEA (p<0.001).

20% of the operations were performed in 67.1% of the hospitals, each of which performed fewer than 10 CEA per annum. A number of hospitals had elevated death rates.

Conclusions

Volume-related improvements in outcome were demonstrated for elective CEA. Minimum volume-criteria of 35 CEA per annum should be established in England. Hospitals performing low annual volumes of surgery should consider arrangements to network services.

Keywords: Carotid endarterectomy, Mortality, Volume, Outcome, Stroke, Centralization

Back to Article Outline

Introduction 

Carotid endarterectomy (CEA) has been shown to be the gold standard in the treatment of symptomatic carotid artery atherosclerotic disease, when there is a significant stenosis.¹, ², ³, ⁴ Following a carotid territory event, appropriately timed CEA can lead to significant reductions in both mortality and morbidity, with studies calculating very low numbers needed to treat (NNT).², ³ However, the absolute risk reduction in adverse outcomes is dependent upon the pathology itself and the surgical risk. Therefore, the best outcomes are partly achieved through the provision of the best surgical care for these patients.

Recent studies have demonstrated a significant volume-related reduction of the in-hospital death rate from high-risk surgical procedures both in North America,⁵, ⁶, ⁷, ⁸, ⁹ and in the UK.¹⁰ A recent meta-analysis, primarily involving data from the USA, of the relationship between volume and outcome in carotid endarterectomy¹¹ suggested that there may be benefits in terms of reduced post-operative stoke and mortality through the limitation of the provision of this procedure to higher-volume hospitals.

The meta-analysis suggested that hospitals should be achieving annual volumes of 70–80 CEA per annum to achieve optimum outcomes. These volumes may not be attainable in England, and so the relationship between surgical volume and outcome should be defined for English practice along with quantification of the critical volume threshold for this procedure. If a relationship is demonstrated between surgical volume and outcome for CEA in England, deaths and adverse neurological events may be prevented.

This study used the English hospital episode statistics (HES) data to investigate the outcome CEA in England from 2000 to 2005. Comparisons were made with current evidence on CEA.

Back to Article Outline

Methods 

HES data were acquired for the years 2000 to 2005. The HES are a data warehouse run by the English Government's Department of Health. The data contain information on every hospital admission in England. An admission (spell) is divided into a number of episodes with an episode being a period of care under one consultant. The data provide patient level information on demographics, the treating hospital, diagnostic and procedural coding, length of stay and in-patient mortality. Diagnostic coding is recorded using the International Classification of Diseases version 10 (ICD-10) diagnostic codes and procedural coding using the Office of Population, Census and Surveys version 4 (OPCS-4) procedural codes.

An in-house program was designed to extract all of the data pertaining to CEA from the database. The data extraction process detected all episodes for all spells between the 1st April 2000 and the 31st March 2005 in which there was an OPCS4 procedure code of interest.

The process was dual stage, with an initial trawl that identified all episodes that matched the criteria and a second run to exclude any episodes that did not have the diagnoses or procedures of interest. This improved the specificity of the data extraction following an initial sensitive search. A further check was run to ensure that each record was reduced to one record per spell to avoid double counting.

Once extracted, the data were organised into elective and emergency clinical groups and cleaned using a further in-house program using the inclusion and exclusion criteria. The clinical groups were determined through the mode of admission (ADMIMETH) specified in HES. Elective admissions were defined as those episodes with an ADMIMETH of 11–13 and emergency admissions were those with an ADMIMETH of 21–24, 28 and 81 (Table 1). OPCS-4 codes L29.4/5/9 were used along with ICD-10 codes I63.0–2, I64, I65.2-3 and I65.8 (Table 2).

Table 1. Summary of Modes of Admission (ADMIMETH) as defined by the HES dictionary. This study utilised ADMIMETH 11-13 to define elective admissions, and ADMIMETH 21-24 and 81 to define emergency admissions

Table 2. Procedural and Diagnostic Codes used in these analyses. OPCS4=Office of Population, Census and Surveys – Procedural Codes version 4. ICD-10=International Classification of Diseases – Diagnostic Codes version 10

Along with the diagnostic and procedural codes, information was extracted regarding the admitting hospital, patient age and gender, length of hospital stay (through admission and discharge dates), timing of operations compared to date of admission. Whether the patient was discharged alive or dead was recorded, and mortality was used as the primary outcome measure. The post-operative stroke rate was investigated, but found not to be a useful measure of outcome (see discussion). Instead, the general complications of surgery were investigated (Table 3).

Table 3. Complications identified from the HES data based on ICD-10 codes

Data quality 

In order to verify the data quality the HES data for the publishing unit (SGH) were compared to the hospitals Patient Administration System (PAS) information over the five-year period in order to highlight any discrepancies. Overall, our centre had a greater than 95% concordance when comparing the databases.

Back to Article Outline

Statistical Analysis 

Summary data were produced for the whole cohort and each clinical group from the verified dataset. Analysis was through evaluation of temporal trends in the data over the five-year period, and assessment of the impact of annual hospital case volume on outcome, with adjustment for age and gender. The dependent variables were in-hospital mortality rate, post-operative complication rate, and length of hospital stay. Elective and emergency admissions were examined independently.

Mortality-control charts 

Mortality-control charts were generated for each clinical group. These showed the mortality rate for individual hospitals against the volume of CEA performed over the five-year period. The national mean mortality rate was shown along with binomial 95% confidence intervals or alarm limits. Centres that lay outside the confidence intervals had a greater than 95% chance that their mortality rate was outside the national average for that procedure.

Volume analysis 

The effect of volume was investigated further through analyses utilising volume quintiles, each containing similar numbers of cases, for each clinical group. The quintiles were arranged to include all hospitals of the same volume in the same quintile, rather than splitting hospitals of the same volume to achieve exactly the same number of cases per quintile. Mortality rates were compared using odds ratios, with the lowest volume quintile in each clinical group set to an odds ratio of one.

Multi-factorial analyses 

For the mortality rate and complication rate, multiple logistic regressions were performed. Maximum likelihood estimates were generated and tested using Chi-squared tests. The independent variables were quantified in terms of proportional odds ratios and 95% Wald confidence limits. Through this methodology, the models predicted estimates of the effect of the year of procedure, and volume, on mortality rate and complication rate, adjusted for age and gender.

For the length of stay, multiple linear regressions were performed on the logarithm of the length of stay using a general linear model procedure, for each clinical group. The independent variables were age, gender and year of procedure and the dependent variable log length of stay. Type III (orthogonal) sum of squares analyses were tested using the F-distribution and the effect of the independent variables on length of stay was quantified through regression estimates. This provided risk-adjusted estimates for the effect of the year of procedure, and volume, on length of hospital stay.

Linear regressions provided estimates and 95% confidence intervals of the trends for the number of hospitals performing CEA each year and for the age of patients undergoing surgery.

The impact of unilateral (I652) and bilateral (I653) carotid stenoses on mortality was assessed using a chi-squared test, with a 2×2 contingency table, for each quintile, and for each clinical group. As more than 25% of the cells had counts of less than five, Fisher's exact test was used to check the validity of the results.

Hospital safety assessment 

Using methods described recently,¹² an assessment was made of the in-hospital death rate at each hospital. The technique was initially described for infra-renal AAA. The aim of this investigation was to assess whether hospitals could provide evidence of safety in CEA, rather than no evidence of danger. In-patient mortality rates were compared using the relative risk (RR) of mortality at a particular hospital compared to the death rate elsewhere. The data were arranged into three groups; hospitals with a RR between zero and one (0<RR<1; green), hospitals with a relative risk between one and two (1<RR<2; blue), and hospitals with a RR greater than two (2<RR; red).

The probability was calculated that the relative risk of mortality at a given hospital was different to twice that elsewhere, based on the number of cases and the number of deaths. The p-values were displayed on a scale of log₁₀(odds) to distinguish small p-values that differed by orders of magnitude. Odds were used rather than p-values to exploit the fact that log (odds) are equal to zero for p=0.5 and so evidence of safety and danger were shown in different directions on the y-axis.

These log₁₀(odds) values were plotted on the y-axis against the hospital procedural volume along the x-axis to construct ‘safety charts’. Log₁₀(odds) of 1.3, equivalent to one-tailed p-value of 0.05, was indicated by solid horizontal lines on the charts. Hospitals that lay outside the two lines generated a significant weight of evidence that their mortality rate was inconsistent with the threshold value, being either higher or lower. Hospitals that lay within the ‘control bands’ may still have had a RR of mortality greater than, or greater than twice, the national average, but there was insufficient evidence to be able to identify them as ‘safe’ or ‘unsafe.’ Overall, this technique provided three alternative states in to which hospitals fell: evidence of safety; insufficient evidence of safety or danger; evidence of danger.

Back to Article Outline

Results 

Summary data 

Between 1st April 2000 and 31st March 2005 there were 280 081 diagnoses of extra-cranial atherosclerotic arterial disease, on which 18 248 CEA (16 759 elective and 1 489 emergency) were performed. Patients had a mean age of 70.2 years for elective CEA and 70.4 years for emergency procedures. Summary data are presented by quintile (Table 4)

Table 4. Summary results for the different clinical groups by volume quintile (Q1-5 where 1 is the lowest volume and 5 the highest volume quintile). I65.2 and I65.3 refer to ICD diagnostic codes

Mortality rates varied from 0.9% in higher volume hospitals to 1.5% at the lowest-volume hospitals.

Mortality control charts 

A mortality control chart was produced for the elective procedures (Fig. 1). Many hospitals had excellent results with no deaths. However, there were a number of outlying low volume hospitals with mortality rates above the upper 95% confidence interval.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1 

  Mortality-Control Chart – Elective CEA. In-patient death rate (drate) against hospital case volume in five years. The horizontal line represents the national mean mortality rate over five years. Upper and lower 95% confidence intervals from the mean are shown. Each point represents the mortality rate over five-years at an individual hospital. There were several low volume hospitals with mortality rates above the 95% confidence interval and the national mean mortality rate.

The effect of volume 

Increasing annual volumes of elective CEA was associated with reduced mortality rates (odds ratio & 95% confidence interval; 0.898 [0.808–0.999]; p=0.047). Elective CEA at the highest volume hospitals reduced the odds of mortality by 36% compared to the lowest volume hospitals. The in-hospital death rate for elective repairs was associated with volume (p=0.047), with a rate of 1.5% in the lowest-volume quintile (<10 repairs p.a.) to 0.95% in the highest-volume quintile (>52 repairs p.a.). The beneficial effects of volume on outcome were apparent from the second quintile upwards.

There was no association between annual hospital volume and mortality for emergency admissions (0.975 [0.798–1.191]; p=0.8026).

Hospital numbers 

A large number of these procedures were performed in low-volume hospitals. Over the five years, for elective surgery, 20% of cases were performed in hospitals doing less than ten CEA per annum, 40% in hospitals doing 17 or fewer. There was no significant change in the number of hospitals providing these services, by either mode of admission, over the five-years of this study.

When translated in to hospital numbers, over the 5 years, 216 of 322 hospitals (67.1%) performed fewer than ten elective CEA per annum, and 265 of 322 (82.3%) of hospitals performed 17 CEA per annum or fewer.

Complication rate and length of stay 

An increasing annual hospital volume was associated with a decreased length of hospital stay for elective (p<0.0001) and emergency procedures (p<0.0001). No relationship was seen between the complication rate and volume for elective CEA (p=0.275) or emergency CEA (p=0.181).

Risk of uni-/bi-lateral stenoses 

Chi-squared analysis of the risk of mortality conferred by unilateral or bilateral carotid artery stenoses was non-significant for elective (p=0.644) and emergency (p=0.165) CEA. The relative proportion of patients with uni-lateral or bi-lateral stenoses was not dependent upon the annual hospital volume of CEA, nor on the mode of admission. These results were confirmed by Fisher's exact test (p=0.243 and p=0.144 for elective and emergency CEA respectively).

Safety charts 

Safety chart assessment (Fig. 2) demonstrated that, for elective CEA, there were six hospitals where the probability of mortality was significantly greater than twice that elsewhere. This provided significant statistical evidence of elevated mortality for CEA at these hospitals. There were a number of hospitals with statistical evidence of safety. A minimum number of 30 CEA per annum was needed in order to be able to evaluate safety.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 2 

  Safety Chart for Elective CEA comparing the mortality rate at a hospital to twice that elsewhere. Relative risk of mortality less than elsewhere are green, RR mortality 1-2 times greater than elsewhere (blue) and more than twice elsewhere (red). Control bands represent p-values of 0.05. Each point represents the log₁₀odds of the p-value for the relative risk of mortality compared to twice that elsewhere. When a point lies above the upper control band this represents evidence of danger. When a point lies below the lower control band this represents evidence of safety. Points between the control bands provide insufficient evidence of safety or danger. Six hospitals had evidence that their mortality rate was more than twice that elsewhere, and therefore demonstrated ‘danger.’ Some hospitals provided evidence of safety based on the number of cases and number of deaths. These hospitals are shown in green and lie below the lower control limit. A minimum of 150 cases in five years (30 per annum) were required in order to demonstrate safety.

Age and gender 

Increasing age had a detrimental effect on the death rate, complication rate, and length of stay for elective CEA. Male gender was associated with an increased length of hospital stay for elective repairs (p<0.001).

Year of procedure 

There was no effect of the year of procedure on in-patient death rates for either elective CEA (p=0.907) or emergency CEA (p=0.577). More recent elective CEA did not have a significantly different complication rate (OR +/− 95% CI 0.961 [0.919–1.004]; p=0.073), whereas a greater number of complications were seen in more recent years for emergency CEA (1.122 [1.020–1.235]; p=0.019). The year of procedure had a significant impact on the length of stay for elective procedures, with more recent CEA having a shorter in-patient stay (p<0.0001). There was no change in length of stay over the study period for emergency CEA (p=0.924).

Back to Article Outline

Discussion 

This study, which utilised current English data, demonstrated a strong relationship between the annual volume of CEA a hospital undertook and outcome. All CEA in England between 2000 and 2005 were included. Hospitals that performed higher annual volumes were associated with lower mortality rates and a shorter in-patient length of stay.

The clinical correlation of ‘emergency’ CEA was likely to be those admitted to hospital as an emergency with a symptomatic carotid artery stenosis. However, the numbers were relatively small, and the correlation uncertain. Therefore, the discussion and conclusions in this study have been drawn exclusively on the elective admissions. Additionally, with elective admissions it was only possible to say that the mode of admission was planned and elective, rather then to explicitly state whether the stenoses were symptomatic. It is likely that the elective group contained a combination of symptomatic and asymptomatic stenoses.

The mortality rate of this cohort was in keeping with current evidence,³, ⁴, ¹³ for example ECST where the quoted mortality rate was 0.9% in those patients with a stenosis of 70–99%.³ This concordance in the mortality rate would suggest that the selection criteria established from ECST have been appropriately adopted in England, and that the trial results still accurately reflect current practice. Discussion of the post-operative stroke rate is reserved for later in this discussion.

A recent meta-analysis¹¹ of the volume-outcome relationship for CEA, based largely on data derived from North America, suggested that hospitals wishing to provide CEA should be capable of providing more than 80 CEA per annum, but accepted that every healthcare system would have to assess its own volume threshold based on the proportion of symptomatic to asymptomatic cases undertaken, as well as the total case volume.

Though this study found that there were some English hospitals achieving these very high annual volumes, the value was clearly in excess of that which could be considered a commonly achievable volume in England. This is highlighted by comparing the thresholds for each volume quintiles identified here to that of Birkmeyer et al.¹⁴ The data from the Nationwide In-patient Sample and Medicare claims database identified ‘very low volume hospitals’ and ‘very high volume hospitals’ as those performing fewer than 40 or more than 164 CEA per annum respectively. This contrasts sharply with this study where equivalent groups performed fewer than 10 and more than 52 CEA per annum. In fact, the US thresholds would include 70% of English CEA provision in the ‘very low volume hospitals’ group.

In England, a minimum volume criterion of 35 CEA per annum should be established, based on the current analyses. This figure was based on the fact that the worst outcomes were seen in the lowest volume quintile, in which the hospitals all performed fewer than ten cases per annum.

216 hospitals performed 20% of the cases and had a mortality rate 0.5% higher than the mean mortality rate of the higher volume quintiles. If these lowest volume hospitals were excluded from performing CEA, then approximately 670 cases would need to be shifted to the remaining 106 hospitals per annum, or approximately six additional cases per annum for each of the remaining hospitals. This increase in caseload should be an easily achievable target for the higher volume hospitals. Each remaining hospital would have to perform an average of 31 CEA per annum in order to meet the current demand. Based on the quintile analysis, this would be equivalent all remaining hospitals being in the top two volume quintiles. Furthermore, this shift in services would also mean that all remaining hospitals would achieve the minimum annual volume required in order to interpret safety, 30 CEA per annum, as established from the safety chart.

Hospitals that are unable to meet the volume-criteria of 35 CEA per annum should be encouraged to refer their cases elsewhere as they cannot then provide evidence of safety in this procedure.

It would be possible to extend this model to exclude the lowest two volume quintiles, in which case 57 hospitals would remain and would be required to take on an additional 23 cases per annum, giving a mean annual volume of 59 cases at each hospital per annum. However, there is evidence that in areas of England that do not have an established CEA service, then instead of patients being referred elsewhere, they are simply not offered CEA.¹⁵ As a result, the suggestion for service reconfiguration is based on figures that provide a sufficient number of cases per hospital to interpret safety whilst significantly improving outcomes and providing the greatest number of hospitals for CEA services.

Case-mix 

In terms of case-mix, higher-volume hospitals performed a greater proportion of operations on older, male patients. Increasing age was shown in this study to have a detrimental effect on survival, and male patients had a longer admission than females. There was no difference seen in mortality for those patients presenting with unilateral or bilateral carotid artery stenoses, and these were not dependent on the annual hospital volume of CEA. As such, there was no case for additional risk-adjustment based on these diagnostic codes.

These data would suggest that higher volume hospitals had a more difficult case-mix, and worse outcomes might be expected. Despite this, for elective surgery, higher volume hospitals had a reduced mortality rate and a shorter length of stay. These findings are in keeping with previous studies that demonstrated that volume related differences in outcome persisted even after adjustment for case-mix.¹⁶, ¹⁷

The level of case-mix adjustment used here was similar to the ‘intermediate model’ described by Aylin et al.¹⁸ for risk-adjustment based on HES data. This was shown to have a similar predictive value for mortality as the most complex risk models available from clinical databases and registries. So, although a relatively simple method, its use can be defended as using the most reliable aspects of the HES data and achieving at least comparable accuracy to models comprising physiological data. That is not to say that future studies should not continue to devise more accurate risk-adjustment models.

Stroke 

The index outcomes for CEA are stroke and mortality, but there is no ICD-10 code to identify specifically a post-operative stroke for these data. Clinical coding was performed retrospectively, and patients may have originally attended with an ipsilateral CVA, TIA, or amaurosis fugax. In this study, there was no way in which to discern pre-operative from post-operative symptoms. An investigation was made in to the stroke rate relative to the annual hospital volume of CEA, but a code for ‘stroke’ was identified in 80 to 90% of CEA cases, making it less useful for analyses of this type.

Clearly this was open to a wide range of interpretations at the time of the clinical event, then again in the subsequent coding. Even if a separate code were available for post-operative stroke, the use of it would remain open to clinical interpretation, a problem that has been highlighted in the past where the inconsistent use of ‘ipsilateral stroke’, ‘disabling stroke’ and ‘any stroke’ have led to difficulties in interpreting CEA outcomes data. These criticisms extend to the large trials upon which CEA selection criteria are based.³, ⁴ With these clinical and coding issues in mind, it was suggested that “the most valid measure for CEA outcome is post-operative death”,¹¹ and the conclusions in this study have been drawn on this basis.

Even if it is accepted that no real information can be drawn on the post-operative stroke rate from these data, the meta-analysis¹¹ did find a correlation between the post-operative mortality rate and stroke rate and it was suggested that mortality could be used as an accurate surrogate for other outcomes of CEA in absence of accurate stroke data. Therefore, based on this statement, and these findings, it is likely that in England, increasing hospital annual volumes of CEA would reduce the post-operative stroke risk, and would add weight to the case towards the rationalisation of care for this procedure.

The effective working of a multi-disciplinary team (MDT) has been identified as a key factor in improving outcomes in many clinical situations. Higher volume hospitals, especially the largest academic hospitals, have easy access to the required MDT for CEA. This includes the routine assessment of all CEA patients by a neurologist pre-operatively and immediate access to vascular laboratories, with specialist technicians.

The impact of hospital infrastructure and appropriate support as a major component of the volume-outcome relationship become clear when considering the fact that, for CEA, lower-volume surgeons achieved results similar to higher-volume surgeons when operating in higher-volume hospitals.¹⁹ This was shown to act independently to the relationship between surgeon operative volume and outcome.

It should be stressed that individual surgeon annual caseload has been shown to be an important factor in the outcome of CEA.¹⁷, ²⁰, ²¹ The degree to which hospital volume or surgeon volume determine outcome is uncertain. It is likely that surgeon volume plays a major role since CEA is a technically demanding procedure in which the physiological stresses are not as significant as for other vascular procedures, such as AAA repair, which are dependent upon critical care expertise to attain the best outcomes. Surgeon level data were not publicly available from HES and so the analyses presented were confined to the relationship between hospital volume and outcome. The overwhelming evidence from previous studies is that death and stroke rate for CEA are lowest for high-volume vascular specialist surgeons, operating in high-volume hospitals.

Finally, these data, being reliant on clinical coding, though accurate for the mortality, date of admission, date of discharge and primary procedural code, contained no information about the degree of stenosis, about patient physiology, details of pre-operative neurology or operative technique. In order for more accurate assessments to be made in the future, emphasis should be placed on submission to databases, which should be led by senior clinicians.

Back to Article Outline

Conclusions 

Volume-related reductions in mortality rate and length of hospital stay were demonstrated for elective CEA. Volume-criteria should be established in England on the basis of these results. Hospitals should perform a minimum of 35 CEA per annum if they wish to provide this procedure. The large number of hospitals performing a low annual volume of surgery should have their workload moved to larger specialist units.

Back to Article Outline

References 

1. Rothwell PM, Eliasziw M, Gutnikov SA, Fox AJ, Taylor DW, Mayberg MR, et al. Analysis of pooled data from the randomised controlled trials of endarterectomy for symptomatic carotid stenosis. Lancet. 2003;361(9352):107–116
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (380 KB)
   • CrossRef
2. Rothwell PM, Eliasziw M, Gutnikov SA, Warlow CP, Barnett HJ. Endarterectomy for symptomatic carotid stenosis in relation to clinical subgroups and timing of surgery. Lancet. 2004;363(9413):915–924
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (115 KB)
   • CrossRef
3. European Carotid Surgery Trial Collaborators Group . Randomised trial of endarterectomy for recently symptomatic carotid stenosis: final results of the MRC European Carotid Surgery Trial (ECST). Lancet. 1998;351(9113):1379–1387
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (109 KB)
   • CrossRef
4. Ferguson GG, Eliasziw M, Barr HW, Clagett GP, Barnes RW, Wallace MC, et al. The North American Symptomatic Carotid Endarterectomy Trial: surgical results in 1415 patients. Stroke. 1999;30(9):1751–1758
   • View In Article
   • MEDLINE
   • CrossRef
5. Holt PJE, Poloniecki JD, Loftus IM, Gerrard D, Thompson MM. Meta-analysis and systematic review of the relationship between volume and outcome in abdominal aortic aneurysm surgery. Br J Surg. 2007;94(4):395–403
   • View In Article
   • MEDLINE
   • CrossRef
6. Sollano JA, Gelijns AC, Moskowitz AJ, Heitjan DF, Cullinane S, Saha T, et al. Volume-outcome relationships in cardiovascular operations: New York State, 1990-1995. J Thorac Cardiovasc Surg. 1999;117(3):419–428[discussion 428–30]
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (175 KB)
   • CrossRef
7. Wen HC, Tang CH, Lin HC, Tsai CS, Chen CS, Li CY. Association between surgeon and hospital volume in coronary artery bypass graft surgery outcomes: a population-based study. Ann Thorac Surg. 2006;81(3):835–842
   • View In Article
   • CrossRef
8. Rathore SS, Epstein AJ, Volpp KG, Krumholz HM. Hospital coronary artery bypass graft surgery volume and patient mortality, 1998-2000. Ann Surg. 2004;239(1):110–117
   • View In Article
   • MEDLINE
   • CrossRef
9. McCabe JE, Jibawi A, Javle P. Defining the minimum hospital case-load to achieve optimum outcomes in radical cystectomy. BJU Int. 2005;96(6):806–810
   • View In Article
   • MEDLINE
   • CrossRef
10. Holt PJE, Poloniecki JD, Loftus IM, Michaels JA, Thompson MM. An Epidemiological study of the relationship between annual surgical volume and outcomes from abdominal aortic aneurysm surgery in the UK from 2000 to 2005. Br J Surg. 2007;94(4):441–448
    • View In Article
    • MEDLINE
    • CrossRef
11. Holt PJE, Poloniecki JD, Loftus IM, Thompson MM. Meta-analysis and systematic review of the relationship between volume and outcome in carotid endarterecomy. Eur J Vasc Endovasc Surg. 2007;33(6):645–651
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (307 KB)
    • CrossRef
12. Holt PJE, Poloniecki JD, Loftus IM, Thompson MM. Demonstrating safety in major surgery: in-hospital mortality following elective repair of abdominal aortic aneurysm in England. Br J Surg. 2007;in press
    • View In Article
13. Rothwell PM, Warlow CP. Prediction of benefit from carotid endarterectomy in individual patients: a risk-modelling study. European Carotid Surgery Trialists' Collaborative Group. Lancet. 1999;353(9170):2105–2110
    • View In Article
    • Full Text
    • Full-Text PDF (119 KB)
    • CrossRef
14. Birkmeyer JD, Siewers AE, Finlayson EV, Stukel TA, Lucas FL, Batista I, et al. Hospital volume and surgical mortality in the United States. N Engl J Med. 2002;346(15):1128–1137
    • View In Article
    • CrossRef
15. Shackley P, Slack R, Michaels JA. Vascular patients' preference for local treatment: an application of conjoint analysis. J Health Serv Res Policy. 2001;6(3):151–157
    • View In Article
    • MEDLINE
    • CrossRef
16. Wennberg DE, Lucas FL, Birkmeyer JD, Bredenberg CE, Fisher ES. Variation in carotid endarterectomy mortality in the Medicare population: trial hospitals, volume, and patient characteristics. JAMA. 1998;279(16):1278–1281
    • View In Article
    • MEDLINE
    • CrossRef
17. Hannan EL, Popp AJ, Tranmer B, Fuestel P, Waldman J, Shah D. Relationship between provider volume and mortality for carotid endarterectomies in New York state. Stroke. 1998;29(11):2292–2297
    • View In Article
    • MEDLINE
    • CrossRef
18. Aylin P, Lees T, Baker S, Prytherch D, Ashley S. Descriptive study comparing routine hospital administrative data with the Vascular Society of Great Britain and Ireland's National Vascular Database. Eur J Vasc Endovasc Surg. 2007;33(4):461–466
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (91 KB)
19. Cebul RD, Snow RJ, Pine R, Hertzer NR, Norris DG. Indications, outcomes, and provider volumes for carotid endarterectomy. JAMA. 1998;279(16):1282–1287
    • View In Article
    • MEDLINE
    • CrossRef
20. Cowan JA, Dimick JB, Thompson BG, Stanley JC, Upchurch GR. Surgeon volume as an indicator of outcomes after carotid endarterectomy: an effect independent of specialty practice and hospital volume. J Am Coll Surg. 2002;195(6):814–821
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (129 KB)
    • CrossRef
21. Pearce WH, Parker MA, Feinglass J, Ujiki M, Manheim LM. The importance of surgeon volume and training in outcomes for vascular surgical procedures. J Vasc Surg. 1999;29(5):768–776[discussion 777–778]
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (83 KB)
    • CrossRef