Impact of Hospital Volume on Chest Tube Duration, Length of Stay, and Mortality After Lobectomy

Transcription

1 Impact of Hospital Volume on Chest Tube Duration, Length of Stay, and Mortality After Lobectomy Hiroshi Otake, MD, MBA, Hideo Yasunaga, MD, PhD, Hiromasa Horiguchi, PhD, Noriyuki Matsutani, MD, Shinya Matsuda, MD, PhD, and Kazuhiko Ohe, MD, PhD Departments of Anesthesiology and Surgery, School of Medicine, Teikyo University, and Departments of Health Management and Policy and Medical Informatics and Economics, Graduate School of Medicine, University of Tokyo, Tokyo; and Department of Preventive Medicine and Community Health, University of Occupational and Environmental Health, Fukuoka, Japan Background. Numerous studies have suggested an inverse relationship between hospital volume and shortterm mortality after various major operations. However, the volume-outcome relationship after lung cancer surgery remains controversial. We investigated the effects of hospital volume on various outcomes after lobectomy for lung cancer, including chest tube duration, postoperative length of stay, and in-hospital mortality. Methods. From a total of 5.85 million inpatients in the Japanese Diagnosis Procedure Combination database, we identified 19,831 patients who underwent lobectomy for lung cancer between July and December in 2007 and Patients were divided into low (<24 per year), medium-low (25 to 43), medium-high (44 to 67), or high (>68) hospital-volume groups. Multivariate regression analyses were conducted to analyze the concurrent effects of various factors on postoperative outcomes. Results. Overall in-hospital mortality was 0.69%, and was significantly lower in the high-volume group compared with the low-volume group (0.48% versus 0.94%; odds ratio 0.60; p 0.047). Chest tube removal occurred earlier in the high-volume group than in the low-volume group (mean 4.0 days versus 5.1; p < 0.001). Postoperative length of stay was shorter in the high-volume group than in the low-volume group (mean 11.5 days versus 15.9, p < 0.001). Conclusions. Higher hospital volume was associated with significantly shorter chest tube duration and postoperative length of stay, and lower in-hospital mortality after lobectomy for lung cancer. However, the differences in outcomes between high-volume and low-volume hospitals may be too small to support regionalization of lung cancer operations to high-volume centers. (Ann Thorac Surg 2011;92: ) 2011 by The Society of Thoracic Surgeons Numerous studies have demonstrated relationships between the number of operations performed at particular hospitals and the surgical outcomes of selected surgical procedures. Most studies have assessed shortterm mortality as the outcome, and concluded that surgical volume was inversely related to operative mortality [1 3]. However, evidence regarding the volume-mortality relationship after lung cancer surgery remains incomplete and controversial. Some studies demonstrated an association between hospital volume and operative mortality [3 6], whereas others found no such relationship [7,8]. Furthermore, data on the effects of hospital volume on other outcomes are scarce [8]. In this study, we investigated the relationship between hospital volume and various clinical outcomes, including duration of chest tube drainage, length of stay (LOS) and in-hospital mortality, after lobectomy for nonsmall cell Accepted for publication April 22, Address correspondence to Dr Yasunaga, Department of Health Management and Policy, Graduate School of Medicine, University of Tokyo, Hongo, Bunkyo-ku, Tokyo , Japan. yasunagah-tky@ umin.ac.jp. lung cancer, using a nationally representative inpatient database, the Japanese Diagnosis Procedure Combination (DPC) database. Material and Methods DPC Database The DPC is a case-mix system similar to the diagnosisrelated groups in Medicare in the United States. This patient classification system was launched in 2002 by the Ministry of Health, Labour, and Welfare of Japan, linked with a lump-sum payment system [9, 10]. All the 82 university teaching hospitals are obliged to adopt the DPC system, but adoption by community hospitals is voluntary. A survey of the DPC hospitals is conducted between July 1 and December 31 each year by the DPC Research Group, funded by the Ministry of Health, Labour, and Welfare. Detailed patient data, as well as administrative claims data, are collected for all the inpatients discharged from the participating hospitals between July 1 and December 31. The survey started with 82 teaching hospitals in 2003, and the number of participat by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 1070 OTAKE ET AL Ann Thorac Surg VOLUME-OUTCOME RELATIONSHIP IN LOBECTOMY 2011;92: ing hospitals had increased to 926 in 2007 and to 855 in 2008, with corresponding patient numbers of 2.99 million and 2.86 million in 2007 and 2008, respectively. The number in 2008 (2.86 million) represented approximately 40% of all the inpatient admissions to acute care hospitals in Japan [9, 10]. The database includes data on patient age and sex; diagnoses and comorbidities recorded with text data in the Japanese language and the International Classification of Diseases, 10th Revision (ICD-10) codes; procedures coded with Japanese original codes; and LOS and discharge status. The requirement for informed consent was waived because of the anonymous nature of the data. Study approval was obtained from the Institutional Review Board at the University of Occupational and Environmental Health. Patient Demographics The patient characteristics examined included sex, age, and preoperative comorbidities, including diabetes mellitus (ICD-10 codes, E10 E14), hypertension (I10 I15), chronic lung diseases (J40 J47), cardiovascular diseases (ischemic heart diseases [I20 I25], valve disorders [I34 I37], cardiomyopathy [I42], and heart failure [I50]), chronic renal failure (N18), liver cirrhosis (K74), and cerebrovascular diseases (I60 I69). According to the DPC coding system for surgical procedures, patients were divided into three surgical groups based on the use of open lobectomy, video-assisted thoracoscopic surgery (VATS) without lymphadenectomy, and VATS with lymphadenectomy. Hospital Volume for Lung Cancer Surgery Hospital volume for all lung cancer operations was determined using the unique identifier for each hospital, and categorized into four volume groups (low, mediumlow, medium-high, and high), with approximately equal numbers of patients in each group. Statistical Analyses We performed univariate comparisons of variables using 2 tests or analysis of variance, as appropriate. Specifically, in-hospital mortality was compared between different subcategories using 2 tests. Logistic regression analysis was performed to model the concurrent effects of hospital volume on in-hospital mortality with adjustment for sex, age and comorbidities. We compared the average durations of chest tube drainage and postoperative LOS between low, medium-low, medium-high, and high hospital volume groups and between the three surgical groups, using analyses of variance. Proportional hazard regression analyses were performed to examine factors affecting earlier removal of chest tube and discharge from hospital. The threshold for significance was a p value less than All statistical analyses were conducted using PASW version 18.0 (SPSS, Chicago, IL). Results From a total of 5.85 million inpatients in the DPC database between July and December, 2007 and 2008, we extracted 19,831 patients who underwent open lobectomy or VATS lobectomy for nonsmall-cell lung cancer (9,516 in 2007 and 10,315 in 2008). Regarding hospital volume, patients were split into low ( 24 per year), medium-low (25 to 43), medium-high (44 to 67), or high hospitalvolume ( 68) groups, with similar numbers of patients in each group. The patient characteristics in each hospital volume group are shown in Table 1. Overall, 11,960 patients (60.3%) were male, and 9,432 (47.6%) were aged 70 years or more. Regarding comorbidities, 3,692 (18.6%) had hypertension, 2,810 (14.2%) had chronic lung dis- Table 1. Patient Characteristics in Each Volume Category Characteristics Total Low ( 24/Year) Medium-Low (25 43/Year) Medium-High (44 67/Year) High ( 68/Year) p Value Number of hospitals Number of patients 19,831 5,013 5,127 4,856 4,835 Sex, male, % Age, years, average SD Comorbidities, % Diabetes mellitus Hypertension Liver cirrhosis Chronic renal failure Cardiovascular diseases Cerebrovascular diseases Chronic lung diseases , % Open lobectomy VATS without lymphadenectomy VATS with lymphadenectomy SD standard deviation; VATS video-assisted thoracoscopic surgery.

3 Ann Thorac Surg OTAKE ET AL 2011;92: VOLUME-OUTCOME RELATIONSHIP IN LOBECTOMY 1071 eases, and 2,258 (11.4%) had diabetes mellitus. Open thoracotomies were performed in 5,840 patients (29.4%), and VATS in 13,991 (70.5%). The average age was lower in the high-volume group. The proportions of patients with comorbidities including diabetes mellitus, hypertension, chronic renal failure, cardiovascular diseases, and cerebrovascular diseases were lower in the highvolume group. The low-volume group included a higher proportion of patients with open lobectomy and a lower proportion with VATS lobectomy than the high-volume group. Table 2 shows the in-hospital mortality for each subcategory. Overall in-hospital mortality was 0.69%. Univariate comparisons showed significantly higher inhospital mortality for males, older patients, patients with diabetes mellitus, chronic renal failure, cardiovascular diseases, and cerebrovascular diseases, and for the lowhospital-volume group. Table 3 shows the results of logistic regression analysis for in-hospital mortality. Significant factors associated with increased mortality were age (odds ratio [OR] 5.88; p for patients aged 80 years compared with patients aged 49) and chronic renal failure (OR 4.53; p 0.001). In-hospital mortality was significantly lower in the high-volume hospital group compared with the lowvolume hospital group (0.48% versus 0.94%; OR 0.60; p 0.047). Table 4 shows the duration of chest tube drainage and postoperative LOS in each hospital volume or surgical group. The mean duration of chest tube drainage was shorter in the high-volume group (4.0 days) than in the low-volume group (5.1 days; p 0.001). The mean postoperative LOS was also shorter in the high-volume group (11.5 days) than in the low-volume group (15.9 days; p 0.001). Compared with open lobectomy, VATS was associated with shorter duration of chest tube drainage and postoperative LOS (p 0.001). Table 5 shows the results of the proportional hazard regression models for chest tube removal rate and discharge rate. Compared with the reference low-volume group, chest tube removal was likely to occur earlier in the medium-low (hazard ratio [HR] 1.23; p 0.001), medium-high (HR 1.28; p 0.001), and high (HR 1.31; p 0.001) hospital volume groups; VATS was associated with earlier chest tube removal than open lobectomy. Similar trends were observed between discharge from hospital and hospital volume or type of surgery; discharge was likely to occur earlier in the medium-low (HR 1.33; p 0.001), medium-high (HR 1.39; p 0.001), and high (HR 1.50; p 0.001) hospital volume groups, and VATS was Table 2. In-Hospital Mortality in Each Subcategory Subcategory n In-Hospital Death (%) p Value Total 19, (0.69) Sex Male 11, (1.01) Female 7, (0.20) Age, years 49 1,036 2 (0.19) ,920 7 (0.24) , (0.54) , (0.85) 80 1, (1.51) Comorbidities Diabetes mellitus 2, (1.15) Hypertension 3, (0.60) Liver cirrhosis 75 1 (1.33) Chronic renal failure (4.17) Cardiovascular diseases 1, (1.31) Cerebrovascular diseases (2.16) Chronic lung diseases 2, (0.68) Hospital volume (per year) Low, 24 5, (0.94) Medium-low, , (0.62) Medium-high, , (0.72) High, 68 4, (0.48) Open lobectomy 5, (0.87) VATS without lymphadenectomy 4, (0.60) VATS with lymphadenectomy 9, (0.62) VATS video-assisted thoracoscopic surgery.

4 1072 OTAKE ET AL Ann Thorac Surg VOLUME-OUTCOME RELATIONSHIP IN LOBECTOMY 2011;92: Table 3. Logistic Regression Analysis of In-Hospital Mortality Subcategory OR 95% CI p Value Sex Male 1.00 Female Age, years Comorbidities Diabetes mellitus Chronic renal failure Cardiovascular diseases Cerebrovascular diseases Hospital volume (per year) Low, 24) 1.00 Medium-low, Medium-high, High, Open lobectomy 1.00 VATS without lymphadenectomy VATS with lymphadenectomy CI confidence interval; OR odds ratio; VATS video-assisted thoracoscopic surgery. associated with earlier discharge from hospital than open lobectomy. Comment The results of studies on operative mortality after lobectomy vary widely because of differences in study populations and years, for example, 4.0% to 5.7% of US Medicare patients [3] and 2.3% of patients aged 65 years or more in New York State [4]. In Canada, in-hospital mortality after lobectomy decreased year by year from 3.1% in 1999 to 1.95% in 2007 [6]. In a French study, in-hospital deaths were identified in 2.4% of 7,480 lobectomy patients between 2002 and 2005 [11], whereas Asian studies have shown relatively low mortalities: 1.3% in Taiwan [7], 0.8% in a previous Japanese study [12], and 0.69% in the present study. Although the reason for the low mortality in Japan remains unclear, it is possible that a high proportion of lobectomy patients in Japan have early-stage cancer. A national survey conducted in 2008 showed that, of 27,881 primary lung cancer patients operated on, 45.9% were diagnosed through a cancerscreening program (including chest roentgenogram and computed tomography), 31.5% were serendipitously diagnosed during follow-up for other diseases, and only 15.3% were diagnosed as a result of close examination for their symptoms. The survey results also showed that clinical stages Ia and Ib comprised 52.4% and 20.2% of the 27,881 patients, respectively [12]. Studies investigating the volume-mortality relationship after lung cancer surgery have provided mixed results. Several studies demonstrated a significant relationship between hospital volume and operative mortality [3 6], whereas some recent studies found no such relationship [7, 8]. The present study, however, demonstrated that higher hospital volume was associated with significantly lower in-hospital mortality after lobectomy, being 0.48% and 0.94% in the high-volume and lowvolume groups, respectively, but the magnitude of the difference was only 0.46%. Extrapolation of these values suggests that of 5,013 operations conducted in the lowvolume hospitals, 23 patients deaths could have been prevented if all these patients had undergone surgery in high-volume hospitals. However, it may be impractical to realize this assumption in an actual clinical situation, in light of patients restricted access to high-volume hospitals. Little is known about the impact of hospital volume on operative outcomes other than mortality after lung cancer surgery. Freixinet and coworkers [8] reported that the overall morbidity after thoracotomy was similar in lowand high-volume hospital groups, while Finley and associates [6] reported that higher volume hospitals showed shorter LOS. The current study revealed that postoperative LOS was shortened by an average of 3.7 days (27.7%) in the high-volume group compared with that in the low-volume group. Our study clarifies the association between hospital volume and chest tube duration after lobectomy; chest tube duration was shortened by an average of 1.1 days (21.6%). Length of stay may reflect overall hospital activities, including surgery, anesthesiology, and intensive medical and nursing care, which could be defined by hospital volume. Longer duration of chest tube drainage, however, could be associated with prolonged air leakage, hemorrhage, or chyiorrhea, and thus be mainly associated with surgical procedures. Table 4. Chest Tube Duration and Postoperative Length of Stay Subcategory Chest Tube Drainage Mean Days (95% CI) Postoperative Length of Stay Mean Days (95% CI) Hospital volume, per year Low, ( ) 15.9 ( ) Medium-low, ( ) 13.1 ( ) Medium-high, ( ) 12.4 ( ) High, ( ) 11.5 ( ) VATS without 3.0 ( ) 10.8 ( ) lymphadenectomy VATS with 4.4 ( ) 12.8 ( ) lymphadenectomy Open lobectomy 5.2 ( ) 15.1 ( ) CI confidence interval; surgery. VATS video-assisted thoracoscopic

5 Ann Thorac Surg OTAKE ET AL 2011;92: VOLUME-OUTCOME RELATIONSHIP IN LOBECTOMY 1073 Table 5. Proportional Hazard Models for Chest Tube Removal and Discharge From Hospital Chest Tube Removal Discharge From Hospital Subcategory HR 95% CI p Value HR 95% CI p Value Sex Male Female Age, years Comorbidities Diabetes mellitus Chronic renal failure Cardiovascular diseases Cerebrovascular diseases Hospital volume, per year Low, Medium-low, Medium-high, High, Open lobectomy VATS without lymphadenectomy VATS with lymphadenectomy CI confidence interval; HR hazard ratio; VATS video-assisted thoracoscopic surgery. Although the current study demonstrated the relationship between higher lobectomy volume and better outcomes in a nationwide setting, this does not indicate that small services are necessarily worse than large services. Some highly skilled community services and surgeons may perform better than large services in teaching hospitals. It is necessary to continue to strive for full disclosure of data and for subsequent quality improvements based on data review and presentation. The practical implications of a known role of hospital volume in predicting outcomes remain controversial. Regionalization to high-volume centers has been bolstered in several countries on the basis of evidence for volume-outcome relationships [6, 13]. However, the results of the present study suggest that the migration of lobectomy to high-volume centers may have limited effects on reducing overall operative mortality. Although we also demonstrated reductions in chest tube duration and postoperative LOS in high-volume hospitals, the magnitude of the reductions was too small to justify regionalization of lung cancer operations to such hospitals. Conversely, regionalization would incur disadvantages in terms of medical accessibility for patients, especially in rural areas, although nationwide health insurance allows all patients free access to any healthcare provider in Japan. Furthermore, increases in volume could overwhelm the resources of high-volume hospitals, thereby rendering the procedures even less accessible. The results of the present study showed that approximately 50% of lobectomies were performed by VATS. This minimally-invasive procedure is highly recommended in Japan, mainly because of its lower morbidity [14-16]. We also confirmed the superiority of VATS for reducing chest tube duration and postoperative LOS. A large proportion of VATS lobectomies in the current study were combined with lymphadenectomy; a previous Japanese study reported that 67.8% of 11,663 lobectomy patients underwent ND2a lymphadenectomy, and 14.6% and 15.6% underwent ND0 and ND1 lymphadenectomy, respectively [17]. According to the Japanese guidelines for lung cancer treatments, systematic dissection is recommended for the precise diagnosis of cancer stage [18]. However, the justification for systematic lymphadenectomy in early lung cancer remains controversial because its effect on long-term survival remains unclear. Regarding LOS, the applicability of our results to other countries is limited. According to the Organization for Economic Cooperation and Development health data, the national average LOS in acute care hospitals in Japan is 18.8 days, compared with only 5.5 days in the United States. This reflects the difference in the nature of hospital care between the two countries; traditionally, most hospitals in Japan provide both early postoperative care and subsequent nursing care for each patient in a single hospitalization. The longer hospital stay in Japan is partly possible because of less pressure on clinicians from

6 1074 OTAKE ET AL Ann Thorac Surg VOLUME-OUTCOME RELATIONSHIP IN LOBECTOMY 2011;92: insurers for the early discharge of patients. In addition, standard practices regarding lobectomy-patient care differ between Japan and other advanced nations. Specifically, Japanese surgeons do not send patients with prolonged air leaks home with outpatient drainage devices, and patients stay in hospital until their chest tube is removed. Several limitations of the present study should be acknowledged. First, this study was based on an administrative claims database, and diagnoses of comorbidities in such a database may be less well validated than those in planned prospective studies. Second, our data lacked full information on several factors that could potentially affect the outcomes, including cancer stage and smoking status. Lastly, the low participation rate of small hospitals in the DPC system skews the evaluable population. The database lacks data from possibly very low volume hospitals, and the overall mortality may therefore have been underestimated. In conclusion, based on information from a national database, the present study demonstrated that higher hospital volume was associated with shorter chest tube duration and postoperative LOS and lower in-hospital mortality after lobectomy for lung cancer. However, the small differences in outcomes between patients at highand low-volume hospitals suggest that regionalization of lung cancer surgery may be impractical, at least in the current clinical environment in Japan. This study was funded by Grants-in-Aid for Research on Policy Planning and Evaluation from the Ministry of Health, Labour, and Welfare, Japan. References 1. Halm EA, Lee C, Chassin MR. Is volume related to outcome in health care? A systematic review and methodologic critique of the literature. Ann Intern Med 2002;137: Killeen SD, O Sullivan MJ, Coffey JC, Kirwan WO, Redmond HP. Provider volume and outcomes for oncological procedures. Br J Surg 2005;92: Birkmeyer JD, Siewers AE, Finlayson EV, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002;346: Hannan EL, Radzyner M, Rubin D, et al. The influence of hospital and surgeon volume on in-hospital mortality for colectomy, gastrectomy, and lung lobectomy in patients with cancer. Surgery 2002;131: Bach PB, Cramer LD, Schrag D, et al. The influence of hospital volume on survival after resection of lung cancer. N Engl J Med 2001;345: Finley CJ, Bendzsak A, Tomlinson G, Keshavjee S, Urbach DR, Darling GE. The effect of regionalization on outcome in pulmonary lobectomy: a Canadian national study. J Thorac Cardiovasc Surg 2010;140: Lien YC, Huang MT, Lin HC. Association between surgeon and hospital volume and in-hospital fatalities after lung cancer resections: the experience of an Asian country. Ann Thorac Surg 2007;83: Freixinet JL, Julia-Serda G, Rodriguez PM, et al. Hospital volume operative morbidity, mortality and survival in thoracotomy for lung cancer: a Spanish multicenter study of 2994 cases. Eur J Cardiothorac Surg 2006;29: Kuwabara K, Matsuda S, Imanaka Y, et al. Injury severity score, resource use, and outcome for trauma patients within a Japanese administrative database. J Trauma 2010;68: Yasunaga H, Yanaihara H, Fuji K, Horiguchi H, Hashimoto H, Matsuda S. Impact of hospital volume on postoperative complications and in-hospital mortality following renal surgery: data from the Japanese diagnosis procedure combination database. Urology 2010;76: Falcoz PE, Conti M, Brouchet L, et al. The Thoracic Surgery Scoring System (Thoracoscore): risk model for in-hospital death in 15,183 patients requiring thoracic surgery. J Thorac Cardiovasc Surg 2007;133: Sakata R, Fujii Y, Kawano H. Thoracic and cardiovascular surgery in Japan during Annual report by the Japanese Association for Thoracic Surgery, Committee for Scientific Affairs. Gen Thorac Cardiovasc Surg 2010;58: Kim SY, Park JH, Kim SG, et al. Disparities in utilization of high-volume hospitals for cancer surgery: results of Korean population-based study. Ann Surg Oncol 2010;17: Whitson BA, Groth SS, Duval SJ, Swanson SJ, Maddaus MA. Surgery for early-stage nonsmall-cell lung cancer: a systematic review of video-assisted thoracoscopic surgery versus thoracotomy approaches to lobectomy. Ann Thorac Surg 2008;86: Villamizar NR, Darrabie MD, Burfeind WR, et al. Thoracoscopic lobectomy is associated with lower morbidity compared with thoracotomy. J Thorac Cardiovasc Surg 2009;138: Paul S, Altorki NK, Sheng S, et al. Thoracoscopic lobectomy is associated with lower morbidity than open lobectomy: a propensity-matched analysis from the STS database. J Thorac Cardiovasc Surg 2010;139: Sawabata N, Fujii Y, Asamura H, et al. [Nationwide statistics on lung cancer surgery in 2004.] J Jpn Assoc Chest Surg 2011;25: The Japan Lung Cancer Society. [Clinical guidelines for the management of lung cancer 2005.] Tokyo, Japan: Kanehara and Company, INVITED COMMENTARY This article [1] examines pertinent clinical endpoints of chest tube duration, length of stay and mortality after lobectomy for lung cancer, as well as the relationship of these variables to surgical volume. It is interesting, therefore, to note that there are no surgeons identified in this authorship. In addition, the Board status, individual surgeon volumes, and extent of resident participation are not specified. These parameters may have a significant impact on the clinical outcomes being studied. The study utilizes The Japanese Diagnosis Procedure Database which is an important reference source with 5.85 million entries including 21,829 lung cancer resections during 2007 to The authors catalogue and recognize the problems associated with such reviews and we will not reiterate. Similarly, we and others use STS, ACC, CMS, Axis, and other databases to address complex questions related to knowledge of current epidemiology and tabulate the results in our pursuit of best practices. It is beneficial to utilize such an expansive 2011 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur