TY - JOUR
AU - Thompson, M M
AB - Abstract Background This article built on previous work to develop an algorithm for elective abdominal aortic aneurysm (AAA) repair and carotid endarterectomy (CEA), with the aim of improving patient survival by regionalization of services. Vascular procedures were used as an example of specialized surgical services. Methods A model was generated based on a national data set that incorporated the statistical demonstration of procedural safety, hospital annual surgical case volume, and travel distance and time. Elective AAA repair was used to construct a hub-and-spoke model that was tested against CEA. The impact of the model was quantified in terms of mortality rates, and travel distance and time. Results Only 48 vascular hubs were required to provide adequate coverage in England, with the majority of patients travelling for less than 1 h to access inpatient vascular surgery. The model predicted a reduction in the number of deaths from elective surgery for AAA (P < 0·001) and CEA (P = 0·016). Conclusion Adoption of this strategic model may lead to improved outcome after AAA and CEA. It could be used as a model for the regionalization of specialized surgery. The model does not take into account the complexity of providing a comprehensive vascular service in every locality. Introduction Recent studies have demonstrated a relationship between both hospital and surgeon workload (volume) and the outcome of arterial procedures in England1–5. These relationships were most striking for elective abdominal aortic aneurysm (AAA) repair2,3,5, but were also significant for carotid endarterectomy (CEA)1,4. New statistical methods have been developed in which ‘safe’ practice and, conversely, ‘unsafe’ practice can be defined6. The latter study suggested that safety and volume were inter-related, with safe practice being dependent, in part, on annual procedural volume. There remain a large number of providers of vascular surgical services in England2,4. This may not be appropriate in light of the studies defining the relationship between caseload and outcome. A recent review of health service organization in the UK has suggested that it might be appropriate to centralize some emergency and specialist surgical services7, but there are limited data to suggest how this regionalization may be justified and implemented. For all specialist and complex surgical services, patient choice will be informed by outcome data and travel times to access care. The present study used the vascular procedures of AAA repair and CEA in England as examples of complex surgical operations, to investigate whether a model of service reconfiguration based on a regionalized system could be informed by operative outcome data and travel times. The findings and interpretation of the model may be applicable to other specialized surgical services. The investigation into the provision of services for AAA is timely with the recent announcement of a national screening programme8 and contemporary publication of death rates after aneurysm repair in the UK press9. The data and conclusions are specific to English vascular surgical practice, but the general message could apply to any healthcare system or complex intervention. The present study investigated a model designed to minimize the mortality rate after elective AAA and CEA. The model was based on the demonstration of safety and the annual volume of surgery performed in various centres, and was assessed on the number of postoperative deaths and the distance of travel for patients to access elective vascular surgery. Methods The study employed Hospital Episode Statistics (HES) in England from 1 April 2000 to 31 March 2005. Details of the data extraction process and statistical methods used previously have been described in full elsewhere2,4,6. In summary, the data extracted were demographic and outcome data for all AAA and CEA procedures performed in English hospitals during the time interval, and included in-hospital mortality and operative complications. The data collected previously demonstrated that elective aneurysm repair was performed in 242 hospitals that corresponded to 410 HES sites. The number of elective open aneurysm repairs per centre ranged from one to 431 over the 5 years. Mortality rates varied widely; the national average was 7·4 per cent. For both elective AAA and CEA, analysis of the amalgamated data by methods described previously2,4 demonstrated a significant relationship between volume and outcome (P < 0·001), with higher-volume hospitals having better outcomes. The data set identified a large number of organizations (HES sites) that delivered vascular services in England. These did not correspond exactly to physical hospitals, especially when ongoing reconfiguration was accounted for, because hospitals merge sites, close and/or undergo redeployment of services. The codes were therefore amalgamated to allocate to a physical hospital all the episodes from all the relevant HES sites over the 5-year interval. Safety plots The nature of the relationship between volume and outcome was examined by producing ‘safety plots’ that investigated whether the mortality rates at individual hospitals were significantly divergent. The threshold investigated was that of twice the national average mortality (k = 2). As an extension to the previous methodology for safety plots, logistic regression was used to produce risk adjustment, such that the plots were adjusted for age and sex. Through the method of data extraction they were also adjusted indirectly for mode of admission. Aims of the model A model was designed to compare contemporary surgical arrangements (current service configuration), as determined from the HES data over the 5-year interval 2000–2005, with a theoretical model of service provision that involved centralization through a hub-and-spoke network. The stimulus for the study was the suggestion that the national average mortality rate for elective arterial surgery is unsatisfactory in England, even after risk adjustment. Some centres had been able to demonstrate safety for arterial procedures, whereas others had not. The model took account of the demonstration of safety and the annual volume of surgery undertaken. The model outputs were: mortality from elective arterial surgery, transport distances and geographical coverage. The model was constructed around elective AAA repair alone. Referral patterns were designed to take account of the factors above. Once the model for elective AAA repair had been optimized, it was tested against CEA procedures from the same interval in England. This provided a method of validation of the model design, together with an estimate of the predicted outcome after CEA. Transport considerations Commercially available software packages10,11 were used to calculate the transport distances and travel times between centres. Although elective procedures were being investigated, the travel time between hospitals was taken as a surrogate for the travel time from home for patients. Travel times and distances were quantified using median values with interquartile range (i.q.r.), as the data were not normally distributed. The travel times and distances were calculated using average times for individual road segments along the most major roads and the most direct route (N. Philips, personal communication). Statistical analysis Calculations were made between each layer of the model to estimate the projected mortality rate and to compare it with that in the current service configuration. Comparisons between groups were made using χ2 analyses and were presented as odds ratio (OR) with 95 per cent confidence interval (c.i.). Assuming that all transferred patients were treated actively, the number needed to treat (NNT) to prevent one death between different model configurations was calculated. Vascular service provision model design Stage 1: safety only Risk-adjusted safety plots were produced for elective AAA repair in the 242 hospitals. Hospitals demonstrating evidence of safety for AAA were identified and nominally assigned as ‘vascular hubs’. Hospitals that did not demonstrate evidence of safety were termed ‘spoke hospitals’ for the purposes of this study. To test the hypothesis, patients were reallocated from the spoke hospitals to the local vascular hub. A meaningful model required that the transferred patients were assumed to adopt the mortality rate of the vascular hub to which they had been transferred. This was a critical assumption, based on the expectation that the mortality rate in the vascular hubs would be lower, as these hospitals had a proven record of safety. The end product of stage 1 was a group of vascular hubs, which had demonstrated safety, undertaking the inpatient elective surgical workload of local spoke hospitals, which had been unable to demonstrate safety. Transport distances and travel times were quantified for all transfers from spoke hospitals into vascular hubs. Stage 2: safety and volume criteria The projected annual volumes of elective AAA repair for stage 1 vascular hubs were quantified. As annual case volume was an independent predictor of outcome from elective aneurysm and carotid surgery in England1–4,6, the model was refined such that all vascular hubs met a minimum volume criterion. Where stage 1 vascular hubs were not predicted to perform an adequate procedure volume (fewer than 32 elective AAA repairs per annum2 and 52 CEAs per annum4), referral patterns were redesigned such that all stage 2 vascular hubs achieved these minimum volumes. Calculations were performed regarding mortality and travel distances, with comparisons being made with the current service configuration and the stage 1 model. Results Stage 1 Risk-adjusted safety plots were produced for elective AAA (Fig. 1). Of 242 hospitals, 70 demonstrated evidence of safety and were designated stage 1 vascular hubs. The remaining patients nationally were reassigned from the spoke hospitals to these vascular hubs. The stage 1 model predicted 830 deaths (5·4 per cent) over 5 years after elective AAA repair compared with 1145 deaths (7·4 per cent) in the current service configuration (OR 0·709 (95 per cent c.i. 0·647 to 0·778); P < 0·001) (Table 1). Fig. 1 Open in new tabDownload slide Risk-adjusted safety plot for elective abdominal aortic aneurysm repair. Adjustment was made for age, sex and mode of admission. Hospitals were as defined in the Hospital Episode Statistics data set and each hospital was represented by a separate symbol. Red symbols represent hospitals with a relative risk of mortality (RR) greater than 2, blue symbols represent hospitals with a RR of 1–2 and green symbols represent hospitals with a RR of less then 1 (better than the national average). The plot shows that some hospitals have demonstrated evidence of safety (green and blue symbols below the dashed lower line), and the remaining hospitals have not. Table 1 Summary data for the model stages compared with the current service configuration . Mortality rate (%) . No. of vascular hubs . No. of deaths . Range of elective AAA repairs per annum . Current service configuration 7·4 242 1145 1·0–73·8 Stage 1 5·4 70 830 6·4–109 Stage 2 5·2 48 807 33·0–132 . Mortality rate (%) . No. of vascular hubs . No. of deaths . Range of elective AAA repairs per annum . Current service configuration 7·4 242 1145 1·0–73·8 Stage 1 5·4 70 830 6·4–109 Stage 2 5·2 48 807 33·0–132 AAA, abdominal aortic aneurysm. Open in new tab Table 1 Summary data for the model stages compared with the current service configuration . Mortality rate (%) . No. of vascular hubs . No. of deaths . Range of elective AAA repairs per annum . Current service configuration 7·4 242 1145 1·0–73·8 Stage 1 5·4 70 830 6·4–109 Stage 2 5·2 48 807 33·0–132 . Mortality rate (%) . No. of vascular hubs . No. of deaths . Range of elective AAA repairs per annum . Current service configuration 7·4 242 1145 1·0–73·8 Stage 1 5·4 70 830 6·4–109 Stage 2 5·2 48 807 33·0–132 AAA, abdominal aortic aneurysm. Open in new tab The NNT was 49·2 (95 per cent c.i. 39·0 to 67·1), meaning that for every 49 elective AAA repairs performed in England in the model one elective death would be prevented over the current service configuration. The median travel time from spoke to hub hospital was 28·5 (range 4–108, i.q.r. 19·0–44·5) min and the distance was 16 (range 0·5–80, i.q.r. 12·0 to 19·0) miles. Stage 2 The 70 stage 1 vascular hubs were assessed for compliance against the minimum volume criteria (more than 32 elective AAA repairs2 and 52 CEAs4 per annum). Twenty-seven of the 70 hubs did not reach this threshold. Referral networks were redesigned, leaving 48 vascular hubs meeting both safety and volume criteria. This revised model predicted 807 deaths over 5 years for elective AAA, which equated to a mortality rate of 5·2 per cent. This was significantly less than in the current service configuration (OR 0·689 (95 per cent c.i. 0·627 to 0·756); P < 0·001). The NNT was 45·9 (95 per cent c.i. 36·9 to 60·8), meaning that for every 46 elective AAA repairs performed in England in the model one elective death would be prevented over the current service configuration. Although a greater number of patients were reallocated to a smaller number of stage 2 vascular hubs, the median travel times (28·5 (range 3–108, i.q.r. 19·0–45·0) min) and distances (16 (range 0·5–80, i.q.r. 8·0–29·0) miles) were not significantly changed from those in stage 1. Model validation against carotid endarterectomy Because the stage 2 model predicted the fewest deaths after elective AAA, and travel times were less than 1 h for most patients, it was tested for efficacy and robustness against CEA. Using the same referral network, CEA procedures were modelled according the stage 2 AAA model. This model predicted 133 deaths compared with 175 deaths in the current service configuration (OR 0·785 (95 per cent c.i. 0·605 to 0·950); P = 0·016). Discussion This study demonstrated how, by restructuring complex surgical services, a significant number of deaths might be prevented. The proposed model had definable transfer times and suggested that service reconfiguration by centralization may improve outcomes. Previous work has demonstrated that mortality rates for elective AAA repair and CEA in England were widely variable and in some instances were disappointingly high6. Hospitals delivering higher volumes of surgery had better outcomes in terms of mortality and postoperative complication rates2,4. More recently, novel techniques have made it possible to investigate individual centres and compare them with other centres nationally to determine evidence of safety in surgery6. The present model built on these findings to investigate the implications in terms of service provision. The new model potentially demonstrated a significantly lower mortality rate from elective arterial surgery than with current arrangements. Hospitals in the model demonstrated evidence of safety and met minimum volume criteria. Each hospital in the stage 2 model would serve a population of approximately 1·5 million. Selecting this model did not significantly increase travel times over the model with a greater number of hospitals (stage 1). The implication is that surgeons would be concentrated in fewer vascular hubs, each with a higher annual caseload, which might be associated with better outcomes. Outcomes from elective surgery may be improved further by these specialist centres gaining experience of minimally invasive techniques such as endovascular aneurysm repair (EVAR) that have also been shown to reduce the mortality rate from elective surgery12,13. In particular, improvements in outcome were demonstrated in North America for EVAR in hospitals adhering to minimum volume criteria14. In addition, reducing death after elective aneurysm repair is obviously crucial to the success of a national screening programme. The annual volumes of surgery suggested by this model are currently being attained by a number of larger hospitals and are not unrealistic. Some hubs may currently find it difficult to provide the higher volume of surgery in the initial stages, and the potential capacity of vascular hubs would need investigation before reconfiguration occurred nationally. This model did not take account of other conditions managed by vascular surgeons, such as peripheral arterial disease. There is little literature in this area concerning the effects of centralizing treatment, in part due to the difficulty in assessing outcome, other than measuring leg amputation rates. It was therefore difficult to construct a similar model for peripheral arterial disease, which may limit its applicability. It would be expected, however, that some specialized services, such as management of the diabetic foot, might be treated more effectively in a centralized model. One potential argument against centralization is the suggestion that patients are concerned with factors other than outcome after elective surgery15, and would prefer to be treated in a local hospital. This concern would need to be addressed and quantified in any planned reconfiguration of services through formal public consultation. However, it is unlikely that a new model of care would require patients to travel to a vascular hub for all stages of care. Outpatient consultations, diagnostic imaging and rehabilitation services could be based at local spoke hospitals, with only complex elective surgery and immediate postoperative care being provided at a hub hospital. These same networks would provide a mechanism for covering vascular emergencies at the spoke hospitals. There would need to be workforce planning, with a smaller number of specialized surgeons required to perform a greater number of complex operations. The equilibrium of workforce may take a considerable time to establish even after service reconfiguration, and might be unpopular among existing surgical teams. Ensuring timely access to a vascular surgeon in an emergency must be a consideration in any model design and transfer times of less than 1 h have been suggested as the acceptable upper limit16. Conversely, there is evidence that patients with a ruptured AAA may have longer interhospital transfers with no significant impact on the operative mortality rate17,18. It may benefit patients with a ruptured AAA to be treated by a surgeon who performs a high volume of elective surgery in a hub hospital19 with the potential to offer emergency EVAR20. In the stage 2 model, emergency transfer of patients from spoke hospitals to vascular hubs was possible in 1 h or less in all but two areas. Local providers may need to be more flexible in designing strategic vascular services in these areas of geographical need in England. England is a relatively densely populated country, and the present model may need to be varied in other countries. These results of this study refer particularly to vascular services in England, but they may be extrapolated to inform discussion regarding the provision of specialized surgical services in other healthcare systems. There is substantial evidence supporting the centralization of cancer and major trauma services in England, Europe and North America. Where these services have been centralized, the subsequent impact of reconfiguration has generally been positive21–24. Outcome monitoring is a fundamental component of any major strategic service change25. The present study made significant assumptions about travel, risk adjustment, mortality rates and data quality. Regarding travel, a patient was assumed to live by his or her spoke hospital. This was on the basis that some patients would live closer and others further away. Nationally, owing to a crossover in hub catchments, the average travel distance would be roughly equivalent to travel from a patient's spoke hospital, although some patients would have to travel significantly further. Risk adjustment is gaining importance in public reporting and scrutiny of surgical outcomes. The complexity of the safety plot methodology has allowed the development of adjustment for age, sex and mode of admission, which have been shown to be the greatest predictors of mortality after major arterial surgery. Previous studies have shown a comparable accuracy between complex physiological risk scores, complex administrative risk scores such as the Romano modification of the Charlson score26 and an intermediate score such as that used here27. This model was built around the assumption that patients from spoke hospitals would adopt the mean mortality rate of their respective hub hospital. The mean mortality rate of a centre over a 5-year interval might be considered a reasonable estimate of the true operative mortality rate in that hospital, independent of case mix, with account taken of both process and structural aspects of care. This was supported by the fact that patients were shown to adopt the lower mortality and stroke rates of hub hospitals when low-volume surgeons from spoke hospitals operated within the hubs28. This assumption was strengthened by hub hospitals demonstrating a higher-risk patient population. The HES data were assumed to be of sufficient accuracy over 5 years to build a model of services, and previous work suggested that HES data were locally accurate over the study interval2. Concerns have been raised that the same level of accuracy may not be attributable to HES data on a national basis. To investigate these concerns was beyond the scope of this study and a multicentre review of the accuracy of HES data is needed to address these concerns. This study was limited by the fact that it only examined in-hospital outcomes. Long-term outcomes and utility scores may also be of interest to commissioners. At present these are not available through the HES, but shortly these data will be linked to the Office of National Statistics. It is hoped that this will provide a mechanism by which to investigate longitudinal outcomes and inform health policy on long-term outcome. This study has proposed a pragmatic approach to the provision of vascular services. A significant reduction in mortality on a national scale could be achieved if the vascular community adopted this model of service provision. This theoretical model is, however, limited by the fact that it only examined two elective vascular procedures. Major strategic change in vascular surgery provision (or any other specialized service) would require a review of all the other elements of vascular practice. Shifting the balance of specialized service provision away from spoke hospitals could have deleterious effects on the treatment of other peripheral vascular diseases, and also other allied services such as stroke care and renal access. References 1 Holt PJ , Poloniecki JD, Loftus IM, Thompson MM. Meta-analysis and systematic review of the relationship between hospital volume and outcome following carotid endarterectomy . Eur J Vasc Endovasc Surg 2007 ; 33 : 645 – 651 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Holt PJ , Poloniecki JD, Loftus IM, Michaels JA, Thompson MM. Epidemiological study of the relationship between volume and outcome after abdominal aortic aneurysm surgery in the UK from 2000 to 2005 . Br J Surg 2007 ; 94 : 441 – 448 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Holt PJ , Poloniecki JD, Gerrard D, Loftus IM, Thompson MM. Meta-analysis and systematic review of the relationship between volume and outcome in abdominal aortic aneurysm surgery . Br J Surg 2007 ; 94 : 395 – 403 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Holt PJE , Poloniecki JD, Loftus IM, Thompson MM. The relationship between hospital case volume and outcome from carotid endarterectomy in England from 2000 to 2005 . Eur J Vasc Endovasc Surg 2007 ; 34 : 646 – 654 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Young EL , Holt PJE, Poloniecki JD, Loftus IM, Thompson MM. Meta-analysis and systematic review of the relationship between surgeon annual volume and outcome in elective open abdominal aortic aneurysm repair . J Vasc Surg 2007 ; 46 : 1287 – 1294 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Holt PJE , Poloniecki JD, Loftus IM, Thompson MM. Demonstrating safety through in-hospital mortality following elective abdominal aortic aneurysm repair in England . Br J Surg 2007 ; 95 : 64 – 71 . Google Scholar Crossref Search ADS WorldCat 7 Darzi A . High Quality Care for All: NHS Next Stage Review Final Report . Department of Health : London , 2008 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 8 Ashton HA , Buxton MJ, Day NE, Kim LG, Marteau TM, Scott RA et al. The Multicentre Aneurysm Screening Study (MASS) into the effect of abdominal aortic aneurysm screening on mortality in men: a randomised controlled trial . Lancet 2002 ; 360 : 1531 – 1539 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 9 Laurance J . Revealed: death rates in your local hospital . The Independent 11 July 2008 . Google Scholar OpenURL Placeholder Text WorldCat 10 AA Route Planner . http://www.theaa.com/travelwatch/planner_main.jsp [accessed 1 October 2008]. 11 Multimap . http://www.multimap.com [accessed 1 October 2008]. 12 Greenhalgh RM , Brown LC, Kwong GP, Powell JT, Thompson SG; EVAR trial participants . Comparison of endovascular aneurysm repair with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day operative mortality results: randomised controlled trial . Lancet 2004 ; 364 : 843 – 848 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Dimick JB , Upchurch GR. Endovascular technology, hospital volume, and mortality with abdominal aortic aneurysm surgery . J Vasc Surg 2008 ; 47 : 1150 – 1154 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Brooke BS , Perler BA, Dominici F, Makary MA, Pronovost PJ. Reduction of in-hospital mortality among California hospitals meeting Leapfrog evidence-based standards for abdominal aortic aneurysm repair . J Vasc Surg 2008 ; 47 : 1155 – 1164 . Google Scholar Crossref Search ADS PubMed WorldCat 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 : 151 – 157 . Google Scholar Crossref Search ADS PubMed WorldCat 16 King M . Innovative new vascular programme combines surgery and radiology . Bull R Coll Surg Engl 2008 ; 90 : 53 . Google Scholar Crossref Search ADS WorldCat 17 Hames H , Forbes TL, Harris JR, Lawlor DK, DeRose G, Harris KA. The effect of patient transfer on outcomes after rupture of an abdominal aortic aneurysm . Can J Surg 2007 ; 50 : 43 – 47 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 18 Salhab M , Farmer J, Osman I. Impact of delay on survival in patients with ruptured abdominal aortic aneurysm . Vascular 2006 ; 14 : 38 – 42 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Cho JS , Kim JY, Rhee RY, Gupta N, Marone LK, Dillavou ED et al. Contemporary results of open repair of ruptured abdominal aortoiliac aneurysms: effect of surgeon volume on mortality . J Vasc Surg 2008 ; 48 : 10 – 18 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Mastracci TM , Garrido-Olivares L, Cina CS, Clase CM. Endovascular repair of ruptured abdominal aortic aneurysms: a systematic review and meta-analysis . J Vasc Surg 2008 ; 47 : 214 – 221 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Granger SR , Glasgow RE, Battaglia J, Lee RM, Scaife C, Shrieve DC et al. Development of a dedicated hepatopancreaticobiliary program in a university hospital system . J Gastrointest Surg 2005 ; 9 : 891 – 895 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Al-Sarira AA , David G, Willmott S, Slavin JP, Deakin M, Corless DJ. Oesophagectomy practice and outcomes in England . Br J Surg 2007 ; 94 : 585 – 591 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Laukontaus SJ , Aho PS, Pettila V, Alback A, Kantonen I, Railo M et al. Decrease of mortality of ruptured abdominal aortic aneurysm after centralization and in-hospital quality improvement of vascular service . Ann Vasc Surg 2007 ; 21 : 580 – 585 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Simons R , Eliopoulos V, Laflamme D, Brown DR. Impact on process of trauma care delivery 1 year after the introduction of a trauma program in a provincial trauma center . J Trauma 1999 ; 46 : 811 – 816 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Holt PJE , Poloniecki JD, Thompson MM. How to improve surgical outcomes . BMJ 2008 ; 336 : 900 – 901 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Romano PS , Roos LL, Jollis JG. Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: differing perspectives . J Clin Epidemiol 1993 ; 46 : 1075 – 1090 . Google Scholar Crossref Search ADS PubMed WorldCat 27 Aylin P , Bottle A, Majeed A. Use of administrative data or clinical databases as predictors of risk of death in hospital: comparison of models . BMJ 2007 ; 334 : 1044 – 1055 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Cebul RD , Snow RJ, Pine R, Hertzer NR, Norris DG. Indications, outcomes, and provider volumes for carotid endarterectomy . JAMA 1998 ; 279 : 1282 – 1287 . Google Scholar Crossref Search ADS PubMed WorldCat Copyright © 2008 British Journal of Surgery Society Ltd. Published by John Wiley & Sons, Ltd. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) Copyright © 2008 British Journal of Surgery Society Ltd. Published by John Wiley & Sons, Ltd.
TI - Model for the reconfiguration of specialized vascular services
JO - British Journal of Surgery
DO - 10.1002/bjs.6433
DA - 2008-11-07
UR - https://www.deepdyve.com/lp/oxford-university-press/model-for-the-reconfiguration-of-specialized-vascular-services-0e11wYtSCM
SP - 1469
EP - 1474
VL - 95
IS - 12
DP - DeepDyve
ER -