Introduction. Short Communication

Transcription

1 bs_bs_banner International Journal of Urology (2013) 20, doi: /iju Short Communication Comparisons of perioperative outcomes and costs between open and laparoscopic radical : A propensity-score matching analysis based on the Japanese Diagnosis Procedure Combination database Toru Sugihara, 1,2 Hideo Yasunaga, 3 Hiromasa Horiguchi, 3 Nobuo Tsuru, 1 Hiroyuki Ihara, 1 Tetsuya Fujimura, 2 Hiroaki Nishimatsu, 2 Kazuhiko Ohe, 4 Kiyohide Fushimi 5 and Yukio Homma 2 1 Department of Urology, Shintoshi Hospital, Iwata, 2 Department of Urology, and Departments of 3 Health Management and Policy and 4 Medical Informatics and Economics, Graduate School of Medicine, The University of Tokyo, and 5 Department of Health Care Informatics, Tokyo Medical and Dental University, Tokyo, Japan Abbreviations & Acronyms CCI = Charlson Comorbidity Index DIC = disseminated intravascular coagulation DPC = Diagnosis Procedure Combination ICD-10 = International Statistical Classification of Diseases and Related Health Problems, Tenth Revision IQR = interquartile range LRP = laparoscopic radical ORP = open radical RALP = robot-assisted laparoscopic radical RP = radical Correspondence: Toru Sugihara M.D., Department of Urology, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo , Japan. ezy04707@nifty.com Received 28 September 2012; accepted 16 December Online publication 16 January 2013 Abstract: We compared perioperative outcomes and costs between open and laparoscopic radical for prostate cancer. The Japanese Diagnosis Procedure Combination database, including cases from 2007 to 2010, was used by one-to-one propensity-score matching. The following items were compared: complication rate; homologous and autologous transfusion rate; first cystography day and cystography repeat rate; anesthesia time; postoperative length of stay; and costs. Multivariate analyses were carried out by including age, Charlson Comorbidity Index, T stage, hospital volume and hospital academic status as variables. As a result, among open and 1997 laparoscopic radical prostatectomies, 1627 propensity-score matched pairs were generated. The laparoscopic approach showed a better overall complication rate (3.4% vs 5.0%), homologous transfusion rate (3.3% vs 9.2%), autologous transfusion rate (44.9% vs 79.3%), first cystography day (mean 6th vs 7th day), mean postoperative length of stay (mean 11 vs 13 days), and cost without surgery and anesthesia (mean $7965 vs $9235; all P < 0.001). Anesthesia time was longer (mean 345 vs 285 min) and total cost was higher (mean $ vs $12 356) for the laparoscopic approach (both P < 0.001). The secondary cystography rates were comparable between the groups (18.3% vs 15.7%, P = 0.144). The multivariate analyses showed similar trends. In conclusion, these findings confirm several benefits of laparoscopy over open approach for radical. Key words: complications, costs and cost analysis, epidemiology, laparoscopic surgery,. Introduction Prostate cancer is a global concern, 1 8 and RP is one of the standard treatments for localized prostate cancer. 1,2,5,8,9 The emergence of RALP has driven RP toward minimally-invasive surgery. 2,10 However, whether the beneficial gain of RALP warrants the high costs remains controversial In Asia, the introduction of robotic systems appears to be sparse compared with Western countries. 10 In Japan, LRP still plays an important role as a major tool for minimally-invasive. Despite several publications comparing LRP and ORP, 2 large-scale multicenter analyses are limited, especially with a focus on Asian populations. In the present study, we hypothesized excellent effectiveness of LRP over ORP on a nationwide level, and carried out multicenter comparisons of perioperative outcomes for ORP and LRP using a Japanese administrative database The Japanese Urological Association 349

2 T SUGIHARA ET AL. Methods Data source Data were obtained from the DPC database, which is a Japanese inpatient administrative claims database. 14,15 The database collected approximately 3 million inpatient data per year, accounting for approximately 40% of all acute inpatient admissions in Japan. 16 The data included main diagnoses, pre-existing comorbidities, and complications recoded using ICD-10 codes, as well as procedures coded with Japanese original codes. 17 The present study was designed as a secondary analysis of the administrative claims data. Given the anonymous nature of the data collection process, informed consent was not required. Study approval was obtained from the Institutional Review Boards and the Ethics Committee of The University of Tokyo. Sampling and end-points We selected patients who underwent ORP or LRP (Japanese codes: K843 and K843-2, respectively) between 2007 and 2010 in the DPC database. Patients bearing colorectal cancer (n = 102) were excluded to simplify the identification process of colorectal injury as described herein below. RALP was not included. Six end-points were set as follows: (i) overall complication rate; (ii) homologous and autologous transfusion rate; (iii) first postoperative cystography day (Japanese code: E003.6) and rate of second cystography; (iv) anesthesia time; (v) postoperative length of stay; and (vi) total costs with or without surgery and anesthesia (80 Japanese yen = US$1). Complications were defined as one of the following items: in-hospital mortality; sepsis (ICD-10 codes: A40.x, A41.x); disseminated intravascular coagulation (D65); pulmonary embolism (I26.x); cardiac events (acute coronary events [I21.x 24.x] or heart failure [I50.x]); stroke (I60.x 64); respiratory complications (J12.x 18.x, J93.x, J95.x, J96.x); acute renal failure (N17.x); peritonitis/peritoneal abscess (K65.x); ileus (K56.x, K91.3); urinary tract infection (N10.x, N30.x, N45.x); disruption of operation wound (T81.3); or colorectal injury. Colorectal injury was judged by the Japanese texts (cho-sonsho/senko) or colorectalrepair surgery or colostomy. (Japanese code: K724, K726). Its judgment validity was reviewed by two authors (TS and HY). Other extracted variables were age, comorbidities at admission, T stage, 18 hospital volume (institutional annual RP caseload) and hospital academic status (academic or non-academic). Comorbidities were converted into the CCI based on Quan s protocol. 19 Hospital volume was categorized into three groups according to tertile volumes. Statistical analysis The significant threshold was P < Univariate comparisons were carried out using the c 2 -test and the Mann Whitney U-test, appropriately. Substantial background differences between the ORP and LRP groups were adjusted by one-to-one propensity-score matching by the nearest neighbor rule with a caliper width of 0.6 of the standard deviation. 1,20,21 This matching process generated a randomized experiment-like situation, which enhances the comparability of both groups. The propensity score was calculated using all explanatory variables. After matching, each end-point was assessed by univariate comparisons. Afterwards, multivariate linear and logistic regression analyses were applied to each of the end-points, appropriately including the covariates of age, CCI, T stage, hospital volume and hospital academic status. All statistical analyses were carried out using IBM SPSS version 19.0 (IBM, Armonk, NY, USA). Results Overall, ORP cases from 722 hospitals and 1997 LRP cases from 58 hospitals were identified. Table 1 shows the patient backgrounds in both groups were closely balanced by the propensity-score matching, which resulted in 1627 matched pairs. The C-statistic was Table 2 shows the outcomes after the propensity-score matching. The complications rate, transfusion rate, first cystography day, postoperative length of stay, and costs without surgery and anesthesia were significantly favorable for the LRP group, whereas the LRP group recorded significantly longer anesthesia time and higher total costs than the ORP group. The secondary cystography rates were equivalent. The multivariate regression analyses shown in Table 3 resulted in the almost similar differences from the univariate comparisons in the end-points. Discussion Our analysis showed several beneficial and disadvantageous aspects of LRP. Despite the increase of RALP booming, 1,2 its high running costs would be a major bottleneck for its further spreading Several recent large-scale systematic reviews have shown some superiority of RALP and LRP over ORP; however, that of RALP over LRP remains controversial. For example, Tewari et al. showed that RALP was superior in complication rates and positive surgical margins, whereas conversion and transfusion incidences were similar among RALP and LRP. 2 Other systematic reviews showed significant advantages in favor of RALP over LRP were observed for operation time, transfusion rate and urinary continence, but not regarding complications, positive surgical margin, The Japanese Urological Association

3 Comparison of outcomes Table 1 Characteristics of the patients who underwent ORP or LRP before and after propensity-score matching Before propensity-score matching After propensity-score matching ORP (%) LRP (%) P ORP (%) LRP (%) P Total Age (years) (29.6) 751 (37.6) < (35.2) 595 (36.5) (32.2) 641 (32.1) 551 (33.8) 528 (32.5) (38.2) 605 (30.3) 503 (30.9) 504 (31.0) CCI (79.3) (78.9) (82.9) 1313 (80.7) (14.8) 295 (14.7) 203 (12.5) 221 (13.6) (5.9) 127 (6.4) 75 (4.6) 93 (5.7) T category T (46.4) (57.3) < (54.9) 935 (57.5) T (8.1) 86 (4.3) 83 (5.1) 79 (4.9) Tx (45.5) 766 (38.4) 651 (40.0) 613 (37.6) Hospital volume Low: 17/year (38.2) 102 (5.1) < (6.2) 102 (6.3) Medium: 18 36/year (33.5) 451 (22.6) 497 (30.5) 447 (27.5) High: 37/year (28.3) (72.3) (63.3) 1078 (66.2) Type of hospital Non-academic (82.7) 543 (27.2) < (33.4) 543 (33.4) Academic (17.3) (72.8) (66.6) 1084 (66.6) Year (25.6) 286 (14.3) < (13.5) 256 (15.7) (22.3) 487 (24.4) 303 (18.6) 327 (20.1) (25.6) 334 (16.7) 308 (18.9) 299 (18.4) (26.5) 890 (44.6) 798 (49.0) 745 (45.8) biochemical recurrence, potency or blood loss. 3,4,6,7 With theses inconsistent results, the cost-effectiveness of RALP compared with LRP would require careful discussion. In Japan, RALP was not covered by the national healthcare system until April 2012, and was sparse, as with other Asian countries. 10 Therefore, LRP still occupies an important role as a minimally-invasive surgery. A notable finding in our analyses was that the two-thirds of complication rates and one-third of homologous transfusion rates in LRP compared with ORP. The low blood loss, partially owing to the pneumoperitoneum pressure, would contribute to the low transfusion use of LRP. 22 Predictably, LRP required approximately 1 h longer anesthesia time, 22,23 but this disadvantage would be justifiable because of the low complication rate. The shorter postoperative stay of LRP would be partially driven by the earlier postoperative cystography and good pain control. 23 Despite the earlier cystography of LRP than ORP, the frequencies of secondary cystography were comparable, as shown in another systematic review. 2 The high laparoscopic privilege fee caused high total costs in the LRP group, but without operation room costs, LRP became preferable to ORP, which could arise by virtue of the shorter stay and lower complication rate. There are other findings that require mention. During the study period, LRP rapidly increased its share from 6.6% in 2007 to 17.8% in Younger patients with early-stage tumors were more likely to receive LRP. Younger patients were more eager for shorter hospitalization and earlier return to work brought about by minimally-invasive surgery. Concerns regarding the extent of invasion and adhesion of advanced tumors could elevate the operation difficulty, and would lead to ORP performance. Biases of LRP toward academic and high-volume institutes were plausible, because hospital volume is a well-known structural indicator of medical quality, and academic hospitals would be keen to introduce new technology. 1,24 Some limitations should be mentioned. First, we failed to include potentially influential information, such as body mass index, surgical details (e.g. lymph node dissection or nerve sparing), past history of abdominal surgery or hormone therapy and pathological details. These unobserved factors could cause systematic biases even after our propensity-based adjustments. Furthermore, data on the positive surgical margin, continence, potency and recurrence-free survival were unavailable in the present study. Second, the analysis of an administrative claims data The Japanese Urological Association 351

4 T SUGIHARA ET AL. Table 2 Comparisons of the propensity-score adjusted outcomes between patients who underwent ORP or LRP After propensity-score matching ORP LRP P n = 1627 n = 1627 Overall complications (%) 81 (5.0) 56 (3.4) Mortality 0 (0.0) 0 (0.0) Sepsis/DIC 1 (0.1) 4 (0.2) Pulmonary embolism 1 (0.1) 0 (0.0) Cardiac events 5 (0.3) 11 (0.7) Stroke 3 (0.2) 2 (0.1) Respiratory complications 6 (0.4) 0 (0.0) Acute renal failure 1 (0.1) 0 (0.0) Peritonitis 5 (0.3) 3 (0.2) Ileus 6 (0.4) 3 (0.2) Urinary tract infection 40 (2.5) 27 (1.7) Wound dehiscence 14 (0.9) 3 (0.2) Colorectal injury 9 (0.6) 8 (0.5) Homologous transfusion (%) 149 (9.2) 53 (3.3) <0.001 Autologous transfusion (%) (79.3) 731 (44.9) <0.001 First cystography day (postoperative day; median, IQR) 7 (6 8) 6 (5 7) <0.001 Second cystography (%) 148 (15.7) 158 (18.3) Anesthesia time, min (median, IQR) 285 ( ) 345 ( ) <0.001 Postoperative length of stay (d; median, IQR) 13 (10 17) 11 (8 13) <0.001 Total costs (US$; median, IQR) ( ) ( ) <0.001 Costs without surgery and anesthesia (US$; median, IQR) ( ) ( ) <0.001 Compared using the Mann Whitney U-test. The c 2 -test was used for other comparisons. Data were available for 940 ORP cases and 862 LRP cases. US$1 = 80 Japanese yen. Table 3 Multivariate regression analyses for propensity-score adjusted outcomes of LRP compared with ORP OR 95% CI P Overall complications (%) to Homologous transfusion (%) to 0.46 <0.001 Autologous transfusion (%) to 0.22 <0.001 First cystography day (day) to -1.5 <0.001 Second cystography (%) to Anesthesia time (min) to 71 <0.001 Postoperative length of stay (day) to -2.5 <0.001 Total costs (US$) +3.1% 2.0 to 4.2 <0.001 Costs without surgery and anesthesia (US$) -12.2% to <0.001 Costs were transformed into log-10 values for the analysis because of their skew distributions. base is intrinsically associated with the problem of miscoding. Third, the costs and hospital stay are affected by the healthcare system. A long hospitalization period in Japan is well known. 25 Finally, bias toward large bed-volume institutes in the DPC database is known. 26 Despite these limitations, our population-based analyses provided worthwhile information that LRP was associated with favorable perioperative outcomes with some acceptable disadvantages compared with ORP. Acknowledgments This study was funded by a Grant-in-Aid for Research on Policy Planning and Evaluation from the Ministry of Health, Labour and Welfare, Japan (No. H22-Policy-031), a Grantin-Aid for Scientific Research B (No ) from the Ministry of Education, Culture, Sports, Science and Technology, and a grant from the Funding Program for World- Leading Innovative R&D on Science and Technology The Japanese Urological Association

5 Comparison of outcomes (FIRST program) from the Council for Science and Technology Policy, Japan (No ). Conflict of interest None declared. References 1 Trinh QD, Sammon J, Sun M et al. Perioperative outcomes of robot-assisted radical compared with open radical : results from the nationwide inpatient sample. Eur. Urol. 2012; 61: Tewari A, Sooriakumaran P, Bloch DA, Seshadri-Kreaden U, Hebert AE, Wiklund P. Positive surgical margin and perioperative complication rates of primary surgical treatments for prostate cancer: a systematic review and meta-analysis comparing retropubic, laparoscopic, and robotic. Eur. Urol. 2012; 62: Novara G, Ficarra V, Rosen RC et al. Systematic review and meta-analysis of perioperative outcomes and complications after robot-assisted radical. Eur. Urol. 2012; 62: Novara G, Ficarra V, Mocellin S et al. Systematic review and meta-analysis of studies reporting oncologic outcome after robot-assisted radical. Eur. Urol. 2012; 62: Jeong CW, Jeong SJ, Hong SK et al. Nomograms to predict the pathological stage of clinically localized prostate cancer in Korean men: comparison with Western predictive tools using decision curve analysis. Int. J. Urol. 2012; 19: Ficarra V, Novara G, Rosen RC et al. Systematic review and meta-analysis of studies reporting urinary continence recovery after robot-assisted radical. Eur. Urol. 2012; 62: Ficarra V, Novara G, Ahlering TE et al. Systematic review and meta-analysis of studies reporting potency rates after robot-assisted radical. Eur. Urol. 2012; 62: Fujimoto H, Nakanishi H, Miki T et al. Oncological outcomes of the prostate cancer patients registered in 2004: report from the Cancer Registration Committee of the JUA. Int. J. Urol. 2011; 18: Takenaka A, Tewari AK. Anatomical basis for carrying out a state-of-the-art radical. Int. J. Urol. 2012; 19: Sundram M. Asian robotic experience. Urol. Oncol. 2010; 28: Hyams ES, Mullins JK, Pierorazio P, Partin AW, Allaf M, Matlaga B. Impact of robotic technique and surgical volume on the cost of radical. J. Endourol. 2012; doi: /end Bolenz C, Gupta A, Hotze T et al. Cost comparison of robotic, laparoscopic, and open radical for prostate cancer. Eur. Urol. 2010; 57: Barbash GI, Glied SA. New technology and health care costs the case of robot-assisted surgery. N. Engl. J. Med. 2010; 363: Sugihara T, Yasunaga H, Horiguchi H et al. Impact of hospital volume and laser use on postoperative complications and in-hospital mortality in cases of benign prostate hyperplasia. J. Urol. 2011; 185: Sugihara T, Yasunaga H, Horiguchi H et al. Renal haemorrhage risk after extracorporeal shockwave lithotripsy: results from the Japanese Diagnosis Procedure Combination Database. BJU Int. 2012; 110: E Sugihara T, Yasunaga H, Horiguchi H et al. Impact of surgical intervention timing on the case fatality rate for Fournier s gangrene: an analysis of 379 cases. BJU Int. 2012; 110: E Sugihara T, Yasunaga H, Horiguchi H et al. Management trends, angioembolization performance and multiorgan injury indicators of renal trauma from Japanese administrative claims database. Int. J. Urol. 2012; 19: Sobin LH, Wittekind C. TNM Classification of Malignant Tumours, 6th edn. Wiley, New York, Quan H, Sundararajan V, Halfon P et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med. Care 2005; 43: Griswold ME, Localio AR, Mulrow C. Propensity score adjustment with multilevel data: setting your sites on decreasing selection bias. Ann. Intern. Med. 2010; 152: Abdollah F, Schmitges J, Sun M et al. Comparison of mortality outcomes after radical versus radiotherapy in patients with localized prostate cancer: a population-based analysis. Int. J. Urol. 2012; 19: Guillonneau B, Rozet F, Cathelineau X et al. Perioperative complications of laparoscopic radical : the Montsouris 3-year experience. J. Urol. 2002; 167: Guazzoni G, Cestari A, Naspro R et al. Intra- and peri-operative outcomes comparing radical retropubic and laparoscopic radical : results from a prospective, randomised, single-surgeon study. Eur. Urol. 2006; 50: Trinh QD, Schmitges J, Sun M et al. Radical at academic versus nonacademic institutions: a population based analysis. J. Urol. 2011; 186: OECD. OECD Health Data Organisation for Economic Co-operation and Development, Paris, France, Sugihara T, Yasunaga H, Horiguchi H et al. Admissions related to interstitial cystitis in Japan: an estimation based on the Japanese Diagnosis Procedure Combination database. Int. J. Urol. 2012; 19: The Japanese Urological Association 353