The impact of certification of general thoracic surgeons on lung cancer mortality: a survey by The Japanese Association for Thoracic Surgery

Transcription

1 European Journal of Cardio-Thoracic Surgery 49 (2016) e134 e140 doi: /ejcts/ezw006 Advance Access publication 31 January 2016 ORIGINAL ARTICLE Cite this article as: Nagayasu T, Sato S, Yamamoto H, Yamasaki N, Tsuchiya T, Matsumoto K et al. The impact of certification of general thoracic surgeons on lung cancer mortality: a survey by The Japanese Association for Thoracic Surgery. Eur J Cardiothorac Surg 2016;49:e134 e140. a The impact of certification of general thoracic surgeons on lung cancer mortality: a survey by The Japanese Association for Thoracic Surgery Takeshi Nagayasu a, *, Shuntaro Sato b, Hiroshi Yamamoto b, Naoya Yamasaki a, Tomoshi Tsuchiya a, Keitaro Matsumoto a, Takuro Miyazaki a, Shunsuke Endo c,d, Fumihiro Tanaka d,e, Hiroyasu Yokomise d,f and Meinoshin Okumura d,g Department of Surgical Oncology, Nagasaki University Graduate School of Biomedical Sciences, Nagasaki, Japan b Clinical Research Center, Nagasaki University Hospital, Nagasaki, Japan c Department of Thoracic Surgery, Jichi Medical University, Tochigi, Japan d Committee for Scientific Affairs, The Japanese Association for Thoracic Surgery, Tokyo, Japan e Second Department of Surgery, University of Occupational and Environmental Health, Fukuoka, Japan f Faculty of Medicine, Department of General Thoracic Surgery, Kagawa University, Kagawa, Japan g Department of General Thoracic Surgery, Osaka University Graduate School of Medicine, Osaka, Japan * Corresponding author. Department of Surgical Oncology, Nagasaki University Graduate School of Biomedical Sciences, Sakamoto, Nagasaki , Japan. Tel: ; fax: ; nagayasu@nagasaki-u.ac.jp (T. Nagayasu). Received 1 September 2015; received in revised form 8 December 2015; accepted 22 December 2015 Abstract OBJECTIVES: The Japanese Board of General Thoracic Surgery and the annual survey by the Japanese Association for Thoracic Surgery ( JATS) of certified hospitals began in 2005; since then, over 1300 specialists and 650 hospitals have been certified by this system. To evaluate how this system contributes to improving the outcomes of general thoracic surgery, the effects of the number of certified general thoracic surgeons (GTSs) and hospital volume on 30-day mortality or hospital mortality were evaluated. METHODS: Using data from the annual survey of JATS from 2005 to 2012, the outcomes of patients who underwent lung resection for lung cancer were evaluated. The patients were divided into four groups by the level of surgery: first level, partial resection; second level, segmentectomy and lobectomy; third level, sleeve segmentectomy and lobectomy; and fourth level, pneumonectomy, sleeve pneumonectomy and pleuro-pneumonectomy. Multiple logistic regression analysis was used to examine the associations between operative mortality and the number of GTSs, hospital volume and level of surgical procedure. RESULTS: Overall 30-day and hospital mortality rates were 0.40 and 0.77%, respectively. The 30-day and hospital mortality rates for each surgical level were 0.20 and 0.35% for the first level, 0.36 and 0.73% for the second level, 1.02 and 1.81% for the third level and 2.42 and 4.26% for the fourth level, respectively. The number of GTSs was associated with lower 30-day and hospital mortality rates (P < ). On logistic analysis, number of GTSs (<3 vs 3), hospital volume (<50 vs 50) and level of procedure (1 vs 2, 3 vs 2, 4 vs 2) were significantly associated with 30-day and hospital mortality rates. For 30-day mortality, the odds ratios were (P < ) for higher number of GTSs and (P = ) for higher volume hospitals. In the subgroup analysis by surgical level, low 30-day and hospital mortality rates in the second and fourth surgical levels were correlated with a higher number of GTSs. CONCLUSIONS: The current decrease in overall 30-day mortality rates from the JATS data showed greater dependence on the number of GTSs than on the hospital volume. We believe that the certification system in Japan is useful for the establishment of GTS status. Keywords: General thoracic surgeons Lung cancer Mortality Hospital volume The Japanese Association for Thoracic Surgery Surgical level INTRODUCTION The Japanese Board of General Thoracic Surgery, organized by the members of the Japanese Association for Thoracic Surgery ( JATS) and the Japanese Association for Chest Surgery ( JACS), was Presented at the 29th Annual Meeting of the European Association for Cardio- Thoracic Surgery, Amsterdam, Netherlands, 3 7 October founded in 2005; to date, over 1300 specialists and 650 hospitals have been certified by this system. The annual nationwide surveys of thoracic surgery by JATS were started in 1986, but until 2004, the Board of General Thoracic Surgery was independently approved by JACS and JATS. After the new system started in 2005, more than items of data have been accumulated by the end of 2012, maintaining high collection rates [1 8]. The Author Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.

2 T. Nagayasu et al. / European Journal of Cardio-Thoracic Surgery e135 This institutional-based survey was replaced in 2015 by the National Clinical Database (NCD), which began collecting data from individual patients in At the time of transition, it was necessary to evaluate how the current system of the Japanese Board of General Thoracic Surgery has contributed to improving the outcomes of general thoracic surgery using the data from 2005 to To this end, the impacts of the number of certified general thoracic surgeons (GTSs) and hospital volume on 30-day and hospital mortality rates of pulmonary resection for lung cancer were evaluated by surgical procedure. PATIENTS AND METHODS The review board of JATS approved the electronic database for the annual survey of JATS from 2005 to 2012 used for this study. During the period of data collection, patients underwent general thoracic surgery, and the numbers of certified GTSs increased from 792 in 2005 to 1166 in 2012 [1 8]. In each year, the committee for scientific affairs of JATS sent survey questionnaire forms to all hospitals nationwide in early April, and the data were collected until the end of December. Only hospitals performing more than an average of 25 general thoracic surgery procedures per year during the most recent 3 years participated. During this period, the rate of collection has been maintained between 91 and 97% [1 8]. Thirty-day and hospital mortality rates of patients who underwent lung resection for lung cancer were evaluated. The definitions of the 30-day and hospital mortality rates were based on the published guidelines of the Ad Hoc Liaison Committee for Standardizing Definitions of Prosthetic Heart Valve Morbidity of the Society of Thoracic Surgeons and the American Association for Thoracic Surgery [9]. Thirty-day mortality was defined as death within 30 days of an operation regardless of the patient s geographic location and even if the patient had been discharged from the hospital within those 30 days. Hospital mortality was defined as death within any time interval after an operation if the patient had not been discharged from the hospital. The patients were divided into four groups by the level of surgery based on the operation fee: first level, partial resection; second level, segmentectomy and lobectomy; third level, sleeve segmentectomy and lobectomy; and fourth level, pneumonectomy, sleeve pneumonectomy and pleuro-pneumonectomy. Operative mortality (30-day and hospital) rates were calculated by surgical level and procedure. To evaluate whether operative mortality decreased as a result of the number of GTSs increasing, their association was plotted, and the Cochran-Armitage test was applied. Next, multiple logistic regression was performed to examine the associations of the number of GTSs (<3, 3), hospital volume (<50, 50) and level of surgical procedure with the odds ratio (OR) and 95% confidence interval (95% CI) and the P value of mortality. In addition, subgroup (level of surgical procedure) analyses were performed. A similar analysis of hospital volume was also performed. SAS software, version 9.4 (SAS Institute, Cary, NC, USA) was used for all analyses. All reported P values are two-sided, and they were not adjusted for multiple comparisons. All data were analysed by the Clinical Research Center of Nagasaki University Hospital. RESULTS Surgical levels Overall 30-day and hospital mortality rates were 0.40 and 0.77%, respectively. The 30-day and hospital mortality rates for each surgical level were 0.20 and 0.35% for the first level, 0.36 and 0.73% for the second level, 1.02 and 1.81% for the third level and 2.42 and 4.26% for the fourth level, respectively (Table 1). Number of general thoracic surgeons The 30-day and hospital mortality rates for each number of GTSs were 0.51 and 0.96% for one GTS, 0.41 and 0.80% for two GTSs, 0.32 and 0.60% for three GTSs and 0.29 and 0.64% for four or more GTSs. The number of GTSs was significantly associated with lower 30-day and hospital mortality rates (both P < ) (Table 2 and Fig. 1). Hospital volume The 30-day and hospital mortality rates for each hospital volume were 0.47 and 0.99% for l to <30 per year, 0.50 and 0.96% for 30 Table 1: Surgical level Thirty-day and hospital mortality rates by surgical level Surgical procedure Operation fee a (by video-assisted thoracic surgery) Number of operations 30-day mortality (%) Hospital mortality (%) THORACIC 1 Partial resection ( ) (0.20) 94 (0.35) (0.36) 1280 (0.73) Segmentectomy ( ) (0.21) 65 (0.34) Lobectomy or bilobectomy ( ) (0.38) 1215 (0.77) (1.02) 64 (1.81) Sleeve and wedge segmentectomy (0.39) 3 (1.16) Sleeve and wedge lobectomy or bilobectomy (1.07) 61 (1.86) (2.42) 195 (4.26) Pneumonectomy (2.42) 182 (4.16) Sleeve and wedge pneumonectomy (4.00) 9 (7.20) Pleuro-pneumonectomy (0.00) 4 (5.26) Total (0.40) 1633 (0.77) a Japanese yen.

3 e136 T. Nagayasu et al. / European Journal of Cardio-Thoracic Surgery Table 2: Thirty-day and hospital mortality rates by number of GTSs Number of GTSs All P a Total number of hospitals Total number of operations day mortality (%) 841 (0.40) 329 (0.51) 240 (0.41) 143 (0.32) 129 (0.29) < Hospital mortality (%) 1633 (0.77) 623 (0.96) 464 (0.80) 263 (0.60) 283 (0.64) < GTSs: general thoracic surgeons. a Cochran-Armitage test. Figure 1: Thirty-day and hospital mortality rates by number of GTSs. The number of GTSs is significantly associated with lower 30-day and hospital mortality rates (both P < ). Hospital mortality reaches a plateau between three and more or more GTSs. GTSs: general thoracic surgeons. to <50 per year, 0.41 and 0.77% for 50 to <100 per year and 0.31 and 0.60% for 100 per year. Hospital volume was significantly associated with lower 30-day and hospital mortality rates (both P < ) (Table 3 and Fig. 2). Logistic regression analysis of number of general thoracic surgeons, hospital volume and surgical level On logistic analysis, number of GTSs (<3 vs 3), hospital volume (<50 vs 50) and level of procedure (1 vs 2, 3 vs 2, 4 vs 2) were significantly associated with 30-day and hospital mortality rates. For 30-day and hospital mortality rates, the ORs were (95% CI: , P < ) and (95% CI: , P < ) for a higher number of GTSs and (95% CI: , P = ) and (95% CI: , P < ) for higher volume hospitals, respectively. In comparing each surgical level for 30-day and hospital mortality rates, the ORs versus surgical level 2 were 0.54 (95% CI: , P < ) and 0.47 (95% CI: , P < ) for surgical level 1, (95% CI: , P < ) and (95% CI: , P < ) for surgical level 3 and (95% CI: , P < ) and (95% CI: , P < ) for surgical level 4, respectively (Fig. 3). Subgroup analysis of odds ratios of number of general thoracic surgeons by surgical level In the subgroup analysis by surgical level, the ORs of the number of GTSs for 30-day mortality were (95% CI: ,

4 T. Nagayasu et al. / European Journal of Cardio-Thoracic Surgery e137 Table 3: Thirty-day and hospital mortality rates by hospital volume Hospital volume All <30 30, <50 50, < P a Total number of hospitals Total number of operations day mortality (%) 841 (0.40) 117 (0.47) 173 (0.50) 334 (0.41) 217 (0.31) < Hospital mortality (%) 1633 (0.77) 246 (0.99) 333 (0.96) 636 (0.77) 418 (0.60) < a Cochran-Armitage test. Figure 2: Thirty-day and hospital mortality rates by hospital volume. Hospital volume is significantly associated with lower 30-day and hospital mortality rates (P = and P = ). THORACIC P < ) for surgical level 2 and (95% CI: , P = ) for surgical level 4. The ORs of the number of GTSs for hospital mortality were (95% CI: , P = ) for surgical level 1, (95% CI: , P < ) for surgical level 2 and (95% CI: , P = ) for surgical level 4 (Fig. 4). DISCUSSION In the last decade, many reports regarding the effects of surgeon specialty, surgeon volume and hospital volume on outcomes in general thoracic surgery, especially in lung cancer surgery, have been published, mostly from the USA. Although the volume and target population of the databases used varied, most of them demonstrated that lower mortality was correlated with thoracic surgeon specialty, higher surgeon volume and higher hospital volume [10 19]. However, the background of these reports was different from the present study. For instance, thoracic surgeons in the USA improved short-term and long-term outcomes compared with general surgeons, but general surgeons performed more pulmonary resections than thoracic surgeons and the surgical specialists were not always board-certified thoracic surgeons [11 15, 17, 19]. In contrast, most pulmonary resections for lung cancer in Japan were performed in an authorized hospital where at least one certified general thoracic surgeon was working. In the present study, data from the nationwide annual survey of thoracic surgery, which

5 e138 T. Nagayasu et al. / European Journal of Cardio-Thoracic Surgery Figure 3: Logistic regression analysis of number of GTSs, hospital volume and surgical level. On logistic analysis, number of GTSs (<3 vs 3), hospital volume (<50 vs 50) and level of procedure (1 vs 2, 3 vs 2, 4 vs 2) are significantly associated with 30-day and hospital mortality rates. For 30-day and hospital mortality rates, the ORs are (95% CI: , P < ) and (95% CI: , P < ) for a higher number of GTSs and (95% CI: , P < ) and (95% CI: , P < ) for higher volume hospitals, respectively. GTSs: general thoracic surgeons; CI: confidence interval; OR: odds ratio. Figure 4: Subgroup analysis of ORs of the number of GTSs by surgical level. In the subgroup analysis by surgical level, the ORs of the number of GTSs for 30-day mortality are (95% CI: , P < ) for surgical level 2 and (95% CI: , P < ) for surgical level 4. The ORs of the number of GTSs for hospital mortality are (95% CI: , P = ) for surgical level 1, (95% CI: , P < ) for surgical level 2 and (95% CI: , P = ) for surgical level 4. GTSs: general thoracic surgeons; CI: confidence interval; OR: odds ratio.

6 T. Nagayasu et al. / European Journal of Cardio-Thoracic Surgery e139 was established by the Japanese Association for Thoracic Surgery in 1986, were used. This database was collected from almost all hospitals certified by the Board of General Thoracic Surgery, which was newly founded by the organized members of JACS and JATS in Over 1300 specialists and 650 hospitals were certified by this system by 2012 [8]. Therefore, the data analysed in the present study were highly restricted to certified GTSs and hospitals compared with other reports. The certified hospitals were divided into reference hospitals and affiliated hospitals. Among the 650 hospitals, the reference hospitals accounted for 43%, and the affiliated hospitals accounted for 57%. Minimum requirements of general thoracic surgery procedures per year during the most recent 3 years were 75 per year for reference hospitals and 25 per year for affiliated hospitals. In every year, the training officers in reference hospitals had the responsibility to report the number of procedures and GTSs in each affiliated hospital, including their reference hospitals. The Japanese Board of General Thoracic Surgery checked these reports and also confirmed the details of operations in each hospital using the JATS data. If the number of procedures or GTSs was below criteria, or there was a defect in the reports, the committee recommended revocation of certification to the hospital through the training officers. When comparing the data of Japan with other countries, the current overall 30-day and hospital mortality rates were very low. The present study showed that the average 30-day mortality rate for patients who had undergone lung cancer surgery was 0.4%. Even if the surgical procedure is limited to more than level 2, the rate was only 0.43% (data not shown). Although there was no trend of improvement in the outcome rate from 2005 to 2012 (Supplementary Data), this rate is much lower than that reported for the 30-day mortality of lung cancer surgery in the USA in 2008 (2.3%) [19] and in the European Society of Thoracic Surgeons database from 2007 to 2014 (3.3%) [20]. The differences among countries regarding the percentage of operations performed by certified GTSs, which was previously described, might have affected these results. Interestingly, the low 30-day mortality rate in authorized hospitals decreased more with both high numbers of GTSs and high hospital volume. On logistic regression analysis, the OR for higher number of GTSs was lower than that for higher volume hospitals. When we focused on issues of multicollinearity in logistic regression analysis, the correlation coefficient between the number of GTSs and hospital volume was 0.585, which was a significant but not strong correlation (data not shown). Thus, the number of GTSs has more impact than hospital volume on the current overall 30-day and hospital mortality rates from the JATS data. This study also investigated how GTSs contributed to the levels of surgical procedures for lung cancer. The subgroup analysis demonstrated that lower 30-day and hospital mortality rates for the second and fourth surgical levels, which included the most standard surgical procedures for lung cancer, were correlated with a higher number of GTSs. Bhamidipati et al. demonstrated that mortality for major pulmonary resections was lowest at hospitals with thoracic residency programmes, and thoracic surgery residency programme hospitals had a more than 30% lower OR for mortality after pneumonectomy than general surgery residency hospitals [19]. The higher number of GTSs might contribute to improving surgical safety by appropriate decision-making regarding surgical indications, more intensive postoperative treatment and avoiding preoperative complications. Because the OR for the fourth level, in which pneumonectomy accounts for a high proportion, was lower than that for the second level, the present results indicate that the current cost of pneumonectomy in Japan is not appropriate if specialty is considered. The data of the JATS survey were gathered by the dedicated cooperation of the certified hospitals. During the observation period, the rate of collection was maintained between 91 and 97%, because each certified hospital is obliged to report its data to maintain certification [1 8]. However, there were limitations to this study because data in the present study were collected per institutional unit, not per individual patient. It was difficult to take into account other factors affecting mortality, such as age, performance status and underlying disease. Since the collection of the postoperative morbidities following lung cancer surgery per institutional unit was started from 2007, it was unfortunate that the data extracted for the present study from the JATS survey did not include these data. This institutional-based survey was replaced in 2015 by the NCD, which began collecting data from individual patients in At the time of transition, this study was the best opportunity to evaluate how the current system by the Japanese Board of General Thoracic Surgery has contributed to improving the outcomes of general thoracic surgery. Although we could not use the risk adjustment methodology in the present study, it will be available when using the new database. To the best of our knowledge, the present study is the first report from Asia to evaluate the impact of certification of GTSs on the mortality of general thoracic surgery. The current decrease in overall 30-day mortality rates from the JATS data showed greater dependence on the number of GTSs than on the hospital volume. We believe that the certification system in Japan is useful for the establishment of GTS status. SUPPLEMENTARY MATERIAL Supplementary material is available at EJCTS online. Conflict of interest: none declared. REFERENCES [1] Ueda Y, Osada H, Osugi H; Japanese Association for Thoracic Surgery Committee for Scientific Affairs. Thoracic and cardiovascular surgery in Japan during Annual report by The Japanese Association for Thoracic Surgery. Gen Thorac Cardiovasc Surg 2007;55: [2] Committee for Scientific AffairsUeda Y, Fujii Y, Udagawa H. Thoracic and cardiovascular surgery in Japan during 2006: annual report by The Japanese Association for Thoracic Surgery. Gen Thorac Cardiovasc Surg 2008;56: [3] Committee for Scientific AffairsUeda Y, Fujii Y, Kuwano H. Thoracic and cardiovascular surgery in Japan during Annual report by The Japanese Association for Thoracic Surgery. Gen Thorac Cardiovasc Surg 2009;57: [4] Sakata R, Fujii Y, Kuwano H. Thoracic and cardiovascular surgery in Japan during 2008: annual report by The Japanese Association for Thoracic Surgery. Gen Thorac Cardiovasc Surg 2010;58: [5] Committee for Scientific AffairsSakata R, Fujii Y, Kuwano H. Thoracic and cardiovascular surgery in Japan during 2009: annual report by The Japanese Association for Thoracic Surgery. Gen Thorac Cardiovasc Surg 2011;59: [6] Kuwano H, Amano J, Yokomise H. Thoracic and cardiovascular surgery in Japan during 2010: annual report by The Japanese Association for Thoracic Surgery. Gen Thorac Cardiovasc Surg 2012;60: [7] Amano J, Kuwano H, Yokomise H. Thoracic and cardiovascular surgery in Japan during 2011: annual report by The Japanese Association for Thoracic Surgery. Gen Thorac Cardiovasc Surg. 2013;61: [8] Committee for Scientific Affairs, The Japanese Association for Thoracic SurgeryMasuda M, Kuwano H, Okumura M, Amano J, Arai H et al. THORACIC

7 e140 T. Nagayasu et al. / European Journal of Cardio-Thoracic Surgery Thoracic and cardiovascular surgery in Japan during 2012: annual report by The Japanese Association for Thoracic Surgery. Gen Thorac Cardiovasc Surg 2014;62: [9] Edmunds LH Jr, Clark RE, Cohn LH, Grunkemeier GL, Miller DC, Weisel RD. Guidelines for reporting morbidity and mortality after cardiac valvular operations. The American Association for Thoracic Surgery, Ad Hoc Liaison Committee for Standardizing Definitions of Prosthetic Heart Valve Morbidity. Ann Thorac Surg 1996;62: [10] Hannan EL, Radzyner M, Rubin D, Dougherty J, Brennan MF. The influence of hospital and surgeon volume on in-hospital mortality for colectomy, gastrectomy, and lung lobectomy in patients with cancer. Surgery 2002;131:6 15. [11] Goodney PP, Lucas FL, Stukel TA, Birkmeyer JD. Surgeon specialty and operative mortality with lung resection. Ann Surg 2005;241: [12] Schipper PH, Diggs BS, Ungerleider RM, Welke KF. The influence of surgeon specialty on outcomes in general thoracic surgery: a national sample 1996 to Ann Thorac Surg 2009;88: [13] Farjah F, Flum DR, Varghese TK Jr, Symons RG, Wood DE. Surgeon specialty and long-term survival after pulmonary resection for lung cancer. Ann Thorac Surg 2009;87: [14] Ellis MC, Diggs BS, Vetto JT, Schipper PH. Intraoperative oncologic staging and outcomes for lung cancer resection vary by surgeon specialty. Ann Thorac Surg 2011;92: [15] Ferraris VA, Saha SP, Davenport DL, Zwischenberger JB. Thoracic surgery in the real world: does surgical specialty affect outcomes in patients having general thoracic operations? Ann Thorac Surg 2012;93: [16] von Meyenfeldt EM, Gooiker GA, van Gijn W, Post PN, van de Velde CJ, Tollenaar RA et al. The relationship between volume or surgeon specialty and outcome in the surgical treatment of lung cancer: a systematic review and meta-analysis. J Thorac Oncol 2012;7: [17] Tieu B, Schipper P. Specialty matters in the treatment of lung cancer. Semin Thorac Cardiovasc Surg 2012;24: [18] Falcoz PE, Puyraveau M, Rivera C, Bernard A, Massard G, Mauny F et al. The impact of hospital and surgeon volume on the 30-day mortality of lung cancer surgery: a nation-based reappraisal. J Thorac Cardiovasc Surg 2014;148: [19] Bhamidipati CM, Stukenborg GJ, Ailawadi G, Lau CL, Kozower BD, Jones DR. Pulmonary resections performed at hospitals with thoracic surgery residency programs have superior outcomes. J Thorac Cardiovasc Surg 2013;145:60 6. [20] Database Report Silver Book 2015-ESTS: pages/files/ests%20201silver_book_full_pef.pdf. APPENDIX. CONFERENCE DISCUSSION Scan to your mobile or go to to search for the presentation on the EACTS library Dr H.J. Ankersmit (Vienna, Austria): These data are very clear: The more you operate, the better you are, and the more certifiedyouare,thebetteryouare.the question is, did this have something to do with the Japanese health system, that you get more referrals, that they concentrate the knowledge into a few hospitals? Dr Nagayasu: What do you mean? Dr Ankersmit: That you specialize the profession and you sort of identify centres where these procedures are allowed to be performed. That would be the logical consequence of what you are reporting. Dr Nagayasu: All the hospitals in Japan are certified by this system. What do you mean? Dr Ankersmit: I mean if the government says that just certified and thoracic surgeons are allowed to operate on the thorax, what is the consequence of your study on the society? Dr Nagayasu: Sorry, I don t know what the consequence means. Dr Ankersmit: Consequence means you are showing that you need lots of volume to do better surgeries. Is that correct? Dr Ankersmit: So basically all the low-volume centres have to be closed down. Now I want to know if this happens in Japan. Dr Nagayasu: Actually, 80% of the hospitals enrolled in this study performed more than 50 cases of thoracic surgery per year, and furthermore, 50% of the hospitals performed more than 100 cases per year. The remaining 20% performed less than 50 cases. Indeed, 25 cases per year, is very small. However, this is, in a sense, a relief measure, because in the hospitals where there is a small amount of volume, only one certified general thoracic surgeon is working in that rural area, as opposed to the general thoracic surgical patients. If we increase the minimum requirements, maybe these hospitals cannot maintain the certification and also the patients cannot receive the optimal support. So in a sense, this is a relief measure. Did this answer your question? Dr Ankersmit: Yes. Dr H. Remmen (Aberdeen, UK): I have a couple of remarks. First, you mentioned that there were 648 hospitals taking part in your study? Dr Remmen: But you excluded the hospitals that were doing less than 25 thoracic procedures? Dr Remmen: This is an interesting study and it does explain what your conclusions are, but obviously the big gap will most likely be between the hospitals you actually studied, compared with the ones you didn t study. Do you have any idea about the hospitals that perform less than 25 thoracic surgical operations per year, what the quality there is and the mortality? The mortality overall looks pretty impressive, I would say, it s very low, although the volumes, if you make a very quick calculation, are pretty low as well over the 648 hospitals you actually investigated. You come to a total of operations. That s what I really don t understand. It says 4700 hospitals. That comes to an average of a little over 50, which is a fairly low number as well. That s the other thing where I got lost a little bit, where you re talking about the 4700 hospitals and before that you were talking about 648 hospitals? Dr Nagayasu: As I mentioned, a low-volume hospital in this study is a very small population. Maybe we need to ignore these data. Even excluding these data, the result has maybe similar outcomes, I think. In Japan now, we are discussing the educational programs. Maybe the issues of small volume have been discussed with the ad hoc committee. Dr Remmen: Let me just try to get something clear here: Initially you said you examined 648 hospitals? Dr Remmen: operations. Dr Nagayasu: No, I mean totally for 8 years. Dr Remmen: Sorry? Dr Nagayasu: Total of 8 years. I mean 4700 hospitals were enrolled through 8 years. So maybe the hospitals enrolled in this study is around 600 per year. Dr Remmen: Thank you very much. Obviously, the message is clear. Economy of scale and quality of scale is, I think, a very barren sort of thing. The background is a little different from other countries.