- Email: [email protected]
The National Cancer Database report on advanced-stage epithelial ovarian cancer: Impact of hospital surgical case volume on overall survival and surgical treatment paradigm
Gynecologic Oncology 118 (2010) 262–267
Contents lists available at ScienceDirect
Gynecologic Oncology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y g y n o
The National Cancer Database report on advanced-stage epithelial ovarian cancer: Impact of hospital surgical case volume on overall survival and surgical treatment paradigm Robert E. Bristow a,⁎, Bryan E. Palis b, Dennis S. Chi c, William A. Cliby d a The Kelly Gynecologic Oncology Service, Departments of Gynecology and Obstetrics and Oncology, The Sidney Kimmel Comprehensive Cancer Center, The Johns Hopkins Medical Institutions, Baltimore, MD, USA b National Cancer Data Base, Commission on Cancer at the American College of Surgeons, Chicago, IL, USA c Gynecology Service Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY, USA d Department of Gynecologic Surgery, Mayo Clinic, Rochester, MN, USA
a r t i c l e
i n f o
Article history: Received 23 April 2010 Available online 22 June 2010 Keywords: Ovarian cancer Surgical volume Survival
a b s t r a c t Objective. To examine the effect of hospital procedure volume and other prognostic variables on overall survival outcome and likelihood of receiving standard recommended care among patients with advancedstage epithelial ovarian cancer. Methods. The National Cancer Data Base (NCDB) was searched for patients undergoing primary treatment for FIGO Stage IIIC/IV epithelial ovarian cancer from 1996 to 2005. The average annual surgical procedure volume was derived for each reporting hospital. Quartile ranking discriminated four groups of hospitals based on annual surgical volume: low (b9), intermediate (9–20), high (21–35), and very high (N 35). Cox proportional hazards modeling was used to determine the impact on overall survival of hospital surgical volume adjusted for treatment, FIGO/AJCC stage, ethnicity, age, payer status, household income, and tumor grade. Binomial multivariate logistic regression modeling was used to assess differences in patient demographic, tumor, and treatment variables between high/very high volume hospitals and low/ intermediate volume hospitals. Results. A total of 45,929 patients were identified. After adjusting for other factors, overall survival was significantly correlated with hospital case volume: very high (reference); high (HR 0.98, 95% CI = 0.92–1.04); intermediate (HR 1.08, 95% CI = 1.01–1.15); and low (HR 1.14, 95% CI = 1.07–1.22). Compared to low and intermediate volume hospitals, patients treated at very high and high-volume hospitals were less likely to receive neo-adjuvant chemotherapy (OR = 0.33, 95% CI = 1.18–1.50) or surgery alone (OR = 0.77, 95% CI = 0.73–0.82) instead of initial surgery and adjuvant chemotherapy. Conclusions. Hospital ovarian cancer surgical volume ≥ 21 cases/year is associated with a higher likelihood of patients with Stage IIIC/IV epithelial ovarian cancer receiving standard treatment (surgery followed by adjuvant chemotherapy). Even after adjusting for treatment paradigm and other factors, hospital volume ≥ 21 cases/year was significantly predictive of improved overall survival outcome. © 2010 Elsevier Inc. All rights reserved.
Introduction The American Cancer Society estimated that 21,550 women in the United States will be diagnosed with ovarian cancer in 2009, and 14,600 women will die of this disease [1]. While advances in novel chemotherapeutic agents and treatment strategies continue to show incremental benefits in survival, recent attention has focused on the potential for improved healthcare outcomes through concentration of
⁎ Corresponding author. Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of California, Irvine - Medical Center, 101 The City Drive, Building 56, Room 260 Orange, CA 92686, USA. Fax: +1 714 456 7754. E-mail address: [email protected] (R.E. Bristow). 0090-8258/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2010.05.025
cancer services. For example, Hillner et al. [2] reported findings of a comprehensive review of the health services literature and found that across all studies, the absolute benefit from care at high-volume centers exceeded the benefit from breakthrough treatments and concluded that efforts to concentrate the initial care for all forms of cancer should be undertaken. A number of population-based studies have demonstrated that the surgical expertise and multidisciplinary care provided to patients with ovarian cancer by high-volume surgeons and high-volume centers is superior when compared to low-volume providers [3–12]. However, other investigators have reported contradictory findings [13–15]. With recent attention focused on improving the quality of cancer care, correlative clinical outcome data for healthcare delivery system characteristics can play an important role in quality improvement efforts. Therefore, the
R.E. Bristow et al. / Gynecologic Oncology 118 (2010) 262–267
potential relationship between hospital ovarian cancer case volume and meaningful clinical outcomes must be clearly defined. The National Cancer Data Base (NCDB) is a joint program of the American College of Surgeons Commission on Cancer and the American Cancer Society and serves to collect data for research and quality improvement. Established in 1985, the NCDB contains data from approximately 20 million patients from more than 1430 participating hospitals, and captures nearly 80% of newly diagnosed cancers in the United States each year. The NCDB collects patient demographics, tumor characteristics, clinical and pathologic TNM stage classification, treatments administered, and survival data [16–20]. The objective of this study was to examine and attempt to quantify the effect of hospital procedure volume on overall survival and processes of care for women with advanced-stage invasive epithelial ovarian cancer within the context of demographic and disease-related prognostic factors using the NCDB. Methods Approval to conduct this study was obtained from the Johns Hopkins Medical Institutions Clinical Research Committee and Joint Commission on Clinical Investigation, and the requirement for informed patient consent was waived (IRB study #NA00037503). Invasive epithelial ovarian patients diagnosed between January 1, 1996, and December 31, 2005, were selected from the NCDB by topography code C56.9 (epithelial histologic types 8010 to 8569; 8940 to 8949; 9000) according to the International Classification of Disease— Oncology 3rd edition (ICD-O-3). Subjects were included if they had an invasive epithelial ovarian malignancy alone or as the first of two or more primary cancers. Patients with borderline ovarian tumors were not included. The study population was further selected to include only those patients with International Federation of Gynecology and Obstetrics (FIGO) and American Joint Committee on Cancer (AJCC) Stage IIIC/IV disease who were managed according to one of the following treatment paradigms: initial surgery followed by adjuvant chemotherapy, initial (neo-adjuvant) chemotherapy followed by surgery, or surgery alone. Surgery was defined as any of the following procedures: local tumor destruction, total removal of a tumor, unilateral oophorectomy, bilateral oophorectomy with omentectomy, debulking/ cytoreductive surgery, pelvic exenteration, or other oophorectomy/surgery type not specified. According to NCDB protocols, hospital cancer registries abstracted cases with the Registry Operations and Data Standards (ROADS) and the Facility Oncology Registry Data Standards (FORDS) manuals, a standardized set of data elements and definitions, using information provided by both patients and hospital medical information systems. Patients received all or part of their first course treatment at the reporting facility. All patients were diagnosed and treated at either community, community comprehensive, or teaching/research hospitals, excluding affiliate and free-standing clinics, pediatric hospitals, and Veterans Affairs facilities. Hospitals were required to have reported data for N7 years between 1996 and 2005. The average annual hospital surgical procedure volume was derived for each reporting hospital based on the total number of patients undergoing operative intervention for all stages of invasive epithelial ovarian cancer and the number of years with reported data. Four groups of surgical volume were identified based on quartile rank; N75th percentile = very high volume (N35 cases/year), 51st– 75th percentile = high volume (21–35 cases/year), 25th–50th percentile = intermediate volume (9–20 cases/year), and b25th percentile = low volume (b9 cases/year). Median household income estimates of patient zip code of residence at the time of diagnosis were obtained by linking the NCDB to US Census data, as patient income is not captured in the NCDB. Five-year overall survival rates were calculated from date of diagnosis to date of death or last follow-up. Patients diagnosed from
263
1996 to 2000 were evaluated for survival outcome, as this group of patients had at minimum 5 years of follow-up data. Survival was estimated by the life table method, and statistical comparisons of survival between hospital volume groups were performed with logrank pairwise comparisons. Cox proportional hazards modeling was used to determine the risk of mortality according to hospital surgical volume adjusted for treatment, pathologic stage, ethnicity, age, payer status, household income, and tumor grade [21]. Hazard ratios (HR) with 95% confidence intervals (95% CI) were generated; HR N 1.0 indicates an increased likelihood of death. Binomial Logistic regression modeling was used to assess patient demographic and tumor characteristics associated with hospital procedure volume categories, adjusted for treatment paradigm, pathologic stage, ethnicity, age, payer status, household income, and tumor grade. Odds ratios (OR) with 95% CIs were generated; OR N 1.0 indicates an increased likelihood of chance. The level of statistical significance was set at p b 0.05. All analyses were generated using SPSS 17.0.1 (SPSS Inc, Chicago, IL). Results At the time of analysis, the NCDB contained 111,557 adult female patients aged ≥18 years at the time of invasive epithelial ovarian cancer diagnosis from 1996 through 2005. Of that total patient pool, 45,929 patients were classified as having pathologic FIGO/AJCC Stage IIIC or IV disease and were treated with either surgery followed by chemotherapy, neo-adjuvant chemotherapy followed by surgery, or surgery alone and were treated at hospitals with ≥7 years of reported data. A total of 1184 hospitals reported sufficient data during the study interval: very high volume, n = 76 (6.4%); high volume, n = 148 (12.5%); intermediate volume, n = 304 (25.7%); and low volume, n = 656 (55.1%). Hospital classification, patient demographics, and tumor characteristics are shown in Table 1. Survival analysis A total of 20,600 patients were available for long-term survival analysis from the 1996–2000 cohort with a minimum of 5-years of follow-up. Log-rank pairwise comparisons revealed that 5-year survival rates for patients treated at low and intermediate volume hospitals (b21 surgeries annually) experienced poorer survival outcomes than patients from high and very high volume hospitals (p b 0.001) (Table 2, Fig. 1). Although very high volume hospitals (N35 surgeries annually) had superior overall survival compared to highvolume hospitals (21–35 surgeries annually), this difference only approached statistical significance (p = 0.07). The 1996–2000 patient cohort was examined by a multivariate Cox regression analysis to determine significant and independent associations between overall survival and demographic and clinicalpathologic variables, including hospital procedure volume. After removing patients with missing values, 12,276 patients were available for analysis. Treatment paradigm, pathologic FIGO/AJCC stage, ethnicity, age at diagnosis, primary payer, household income level, and tumor grade were significantly associated with overall survival (Table 3). The survival outcome of patients treated with either neoadjuvant chemotherapy or surgery alone was significantly worse compared to the treatment paradigm of initial surgery followed by adjuvant chemotherapy. After adjusting for the effects of other factors, there was a statistically significant and independent association between increasing hospital procedure volume and overall survival. Using very high volume hospitals as the referent group, there was no significant difference in overall survival compared to high-volume hospitals. However, below a hospital surgical volume of 21 cases/year, there was a stepwise decrease in overall survival as the hospital procedure volume decreased: intermediate volume HR = 1.08 (95% CI = 1.01 to 1.15), low-volume HR = 1.14 (95% CI = 1.07 to 1.22).
264
R.E. Bristow et al. / Gynecologic Oncology 118 (2010) 262–267
Table 1 Hospital volume, patient demographics, and tumor characteristics for epithelial ovarian cancer Stage IIIC/IV, 1996–2005 (n = 45,929).
Treatment paradigm
AJCC pathologic stage Tumor grade
Ethnicity
Age at diagnosis Primary payer
Household income
Surgery only Neoadjuvant chemotherapy + surgery Surgery + adjuvant chemotherapy Surgery–chemotherapy sequence unknown Stage IIIC Stage IV Well/moderately differentiated Undifferentiated/anaplastic Unknown Non-Hispanic White Non-Hispanic Black Non-Hispanic other Hispanic Unknown Mean (years) Private insurance Managed care Medicaid Medicare/Medicare with supplement Not Insured Unknown N$35,000/year b$34,999/year Unknown
Total
Very high
High
Intermediate
Low
Total
2416 (18.1%) 867 (6.5%) 8365 (62.5%) 1737 (13.0%) 9482 (70.8%) 3903 (29.2%) 2606 (19.5%) 8876 (66.3%) 1903 (14.2%) 10774 (80.5%) 730 (5.5%) 211 (1.6%) 452 (3.4%) 1218 (9.1%) 61.6 2497 (18.7%) 4172 (31.2%) 525 (3.9%) 4653 (34.8%) 468 (3.5%) 1070 (8.0%) 3787 (28.3%) 8755 (65.4%) 843 (6.3%) 13385
3048 (21.8%) 745 (5.3%) 8963 (64.2%) 1203 (8.6%) 9634 (69.0%) 4325 (31.0%) 2913 (20.9%) 9013 (64.6%) 2033 (14.6%) 10916 (78.2%) 764 (5.5%) 330 (2.4%) 518 (3.7%) 1431 (10.3%) 62.2 2668 (19.1%) 4355 (31.2%) 510 (3.7%) 5357 (38.4%) 397 (2.8%) 672 (4.8%) 3733 (26.7%) 9376 (67.2%) 850 (6.1%) 13959
2717 (23.9%) 573 (5.1%) 6602 (58.2%) 1454 (12.8%) 7512 (66.2%) 3834 (33.8%) 2372 (20.9%) 7038 (62.0%) 1936 (17.1%) 8828 (77.8%) 738 (6.5%) 276 (2.4%) 607 (5.3%) 897 (7.9%) 62.7 2044 (18.0%) 3425 (30.2%) 469 (4.1%) 4355 (38.4%) 523 (4.6%) 530 (4.7%) 3084 (27.2%) 7640 (67.3%) 622 (5.5%) 11346
1722 (23.8%) 332 (4.6%) 4360 (60.2%) 825 (11.4%) 4114 (56.8%) 3125 (43.2%) 1547 (21.4%) 4116 (56.9%) 1576 (21.8%) 5778 (79.8%) 391 (5.4%) 149 (2.1%) 275 (3.8%) 646 (8.9%) 64 1344 (18.6%) 1661 (22.9%) 334 (4.6%) 3258 (45.0%) 244 (3.4%) 398 (5.5%) 1486 (20.5%) 5379 (74.3%) 374 (5.2%) 7239
9903 2517 28290 5219 30742 15187 9438 29043 7448 36296 2623 966 1852 4192
Characteristics associated with hospital procedure volume The entire ovarian cancer study population was examined for significant independent associations between hospital procedure volume and demographic and clinical-pathologic variables. After excluding patients with missing values, a total of 27,161 patients were available for study. For the purposes of this analysis, very high and high-volume hospitals were combined into a single category, and intermediate and low-volume hospitals were combined into a single category (Table 4). Compared to patients treated at low/intermediate volume hospitals, patients treated at very high/high-volume hospitals were less likely to be treated with neo-adjuvant chemotherapy (OR = 0.33, 95% CI = 0.18 to 0.50) and with surgery alone (OR = 0.77,
(21.6%) (5.5%) (61.6%) (11.4%) (66.9%) (33.1%) (20.5%) (63.2%) (16.25) (79.0%) (5.7%) (2.1%) (4.0%) (9.1%)
8553 (18.6%) 13613 (29.6%) 1838 (4.0%) 17623 (38.4%) 1632 (3.6%) 2670 (5.8%) 12090 (26.3%) 31150 (67.8%) 2689 (5.9%) 45929
95% CI = 0.73 to 0.82) versus surgery followed by adjuvant chemotherapy. In addition, patients from very high/high surgical volume hospitals were less likely to have pathologic stage IV disease and undifferentiated/anaplastic tumors, be of Hispanic ethnicity, be age N 64 years, and have a lower median household income level. Discussion Concentration of cancer care services has been advocated on the basis of improved clinical outcomes associated with clinicians and centers providing high-volume clinical practices. With regard to ovarian cancer, the Society of Surgical Oncology has provided the following guidelines for surgery: “Surgeons undertaking operations
Table 2 Log-rank pairwise comparisons of observed overall survival rates for epithelial ovarian cancer Stage IIIC/IV, 1996–2000 (n = 17,683).
Very high volume
High volume
Intermediate volume
Low volume
Month of survival
Percent survival rate (95% CI)
0 12 24 36 48 60 0 12 24 36 48 60 0 12 24 36 48 60 0 12 24 36 48 60
100% 82.1% 64.1% 48.8% 37.2% 28.9% 100% 79.3% 61.7% 46.6% 35.9% 28.5% 100% 78.6% 58.4% 43.3% 32.9% 26.1% 100% 75.0% 55.6% 40.8% 31.3% 24.3%
Very high volume
High volume
Intermediate volume
Low volume
(81.2%–83.0%) (63.0%–65.3%) (47.6%–50.0%) (36.0%–38.4%) (27.8%–30.1%)
NA
0.07
b0.001
b 0.001
(78.2%–80.5%) (60.4%–63.1%) (45.2%–48.0%) (34.6%–37.3%) (27.2%–29.8%)
0.07
NA
b0.001
b 0.001
(77.1%–80.0%) (56.7%–60.2%) (41.6%–45.1%) (31.3%–34.6%) (24.5%–27.7%)
0.00
b0.001
NA
0.02
(73.3%–76.7%) (53.7%–57.5%) (38.9%–42.8%) (29.5%–33.1%) (22.6%–26.0%)
b 0.001
b0.001
0.02
NA
R.E. Bristow et al. / Gynecologic Oncology 118 (2010) 262–267
265
Table 4 Comparison (binomial logistic regression model) of intermediate/low-volume hospitals (baseline) and very high/high-volume hospitals according to patient demographics and tumor characteristics for epithelial ovarian cancer Stage IIIC/IV, 1996–2005 (n = 27,161). Characteristic
Fig. 1. Overall survival according to hospital surgical case volume for epithelial ovarian cancer Stage IIIC/IV, 1996–2000 (n = 20,600).
for possible ovarian cancer should have both the necessary technical expertise and a thorough understanding of the management of the disease itself…optimal treatment of this disease requires the skillful and appropriate integration of cancer surgery and chemotherapy, and is best carried out in centers in which an experienced and coordinated multidisciplinary team is available” [22]. The association between subspecialist care by a gynecologic oncologist and superior ovarian cancer outcomes has been well documented [3,23–26]. Recently,
Table 3 Risk factors for mortality for epithelial ovarian cancer Stage IIIC/IV, 1996–2000 (n = 10,641). Risk factor Hospital surgical case volume Very high High Intermediate Low Treatment paradigm Surgery + adjuvant chemotherapy Neoadjuvant chemotherapy + surgery Surgery alone FIGO/AJCC stage group Stage IIIC Stage IV Tumor grade Well/moderately differentiated Undifferentiated/anaplastic Ethnicity Non-Hispanic White Non-Hispanic Black Hispanic Age at diagnosis b40 years 40–64 years N64 years Primary payer Private insurance Managed care Medicaid Medicare/Medicare with supplement Not insured Household income N$35,000/year b$34,999/year
N
Significance level
Hazard ratio (95% CI)
4046 3066 1936 1593
0.26 0.03 0.00
Referent 1.03 (0.98–1.09) 1.08 (1.01–1.15) 1.16 (1.09–1.24)
5570 300 4771
0.00 0.01
Referent 1.20 (1.15–1.26) 1.19 (1.04–1.36)
6994 3647
0.00
Referent 1.42 (1.35–1.48)
2871 7770
0.00
Referent 1.16 (1.10–1.22)
9584 670 387
0.01 0.09
Referent 1.14 (1.04–1.25) 0.89 (0.79–1.02)
383 5141 5117
0.21 0.00
Referent 1.09 (0.95–1.25) 1.43 (1.23–1.66)
2465 2935 4505 353 383
0.90 0.00 0.00 0.01
Referent 1.00 (0.93–1.07) 1.35 (1.19–1.54) 1.15 (1.06–1.26) 1.19 (1.03–1.37)
3030 7611
0.00
Referent 1.09 (1.03–1.15)
Treatment paradigm Surgery + adjuvant chemotherapy Neoadjuvant chemotherapy + surgery Surgery alone FIGO/AJCC stage group Stage IIIC Stage IV Tumor grade Well/moderately differentiated Undifferentiated/anaplastic Ethnicity Non-Hispanic White Non-Hispanic Black Hispanic Age at diagnosis b40 years 40–64 years N64 years Primary payer Private insurance Managed care Medicaid Medicare/Medicare with supplement Not insured Household income N$35,000/year b$34,999/year
N
Significance level
Odds ratio (95% CI)
19213 1473 6475
0.00 0.00
Referent 0.33 (0.18–0.50) 0.77 (0.73–0.82)
18369 8792
0.00
Referent 0.72 (0.69–0.76)
6699 20462
0.00
Referent 0.86 (0.81–0.91)
24292 1744 1125
0.12 0.00
Referent 0.92 (0.83–1.02) 0.81 (0.72–0.92)
883 13685 12593
0.19 0.00
Referent 0.91 (0.78–1.05) 0.78 (0.66–0.92)
5454 8399 1205 11002 1101
0.00 0.64 0.37 0.00
Referent 1.14 (1.06–1.23) 0.97 (0.84–1.11) 1.05 (0.95–1.15) 0.80 (0.70–0.93)
7594 19567
0.00
Referent 0.89 (0.84–0.95)
however, the medical community focused on the potential impact of hospital concentration of ovarian cancer care on patient outcomes. Despite this focus, it has been difficult to quantify the possible benefit from hospital concentration of cancer services as a result of the multifactorial survival determinants for patients with ovarian cancer. Given the complexity of surgical and post-surgical care required for the effective management of patients with advanced-stage ovarian cancer in particular, the primary objective of the current study was to explore a potential positive hospital volume–outcome relationship for this group of patients using the resources of the NCDB. The majority of published population-based studies evaluating hospital procedure volume and ovarian cancer outcomes have come from outside of the United States, and the results have been somewhat contradictory [4–7,10–13,27–29]. For example, Marth et al. [10] reported a prospective study of 1948 patients in Austria. After controlling for other prognostic variables, treatment at a low-volume center (b24 cases/year) was a statistically significant and independent negative prognostic factor for overall survival (HR = 1.38, 95% CI = 1.2–1.7, p b 0.001). Ioka et al. [7] reported on 2450 ovarian cancer patients from the Osaka Cancer Registry. These investigators found that, after controlling for other variables, patients cared for at very low-volume hospitals (b1 case/year) had a 60% increase in the risk of disease-related death compared to high-volume hospitals. Similarly, Wimberger et al. [27] reported on 785 patients from Germany, extracted from the AGO-OVAR Study Group, and found that surgery at a high-volume center was associated with a significant reduction in the risk of disease-related death (HR = 0.77, 95% CI = 0.63–0.94, p = 0.012) after controlling for other variables. Not all studies have found a significant correlation between hospital volume and ovarian cancer survival outcome, however. For example, Vernooj et al. [14] investigated the influence of hospital case volume on ovarian cancer survival among 1077 patients from 18 Dutch hospitals between 1996 and 2003. High-volume hospital (≥13 cases/year) care was associated with a higher likelihood of adequate
266
R.E. Bristow et al. / Gynecologic Oncology 118 (2010) 262–267
staging procedures but was not significantly associated with overall survival. Similarly, Kumpulainen et al. [15] conducted a follow-up study of 275 ovarian cancer patients from the Finish Cancer Registry and found that an annual hospital ovarian cancer operative case volume of ≥10 cases was significantly predictive of complete surgical resection but was not an independent predictor of overall survival. There are two U.S. population-based ovarian cancer studies examining the potential impact of hospital case volume [8,13]. Goff et al. [8] reported an analysis of 10,432 patients from the Health Care Cost and Utilization Project investigating the impact of hospital case volume on stage-appropriate surgical procedures. These authors found that, after controlling for other factors, patients operated upon at a highvolume hospital were more likely to undergo lymphadenectomy and tumor cytoreduction [8]. In 2006, Schrag et al. [13] published the only large U.S. population-based outcome study to examine the impact of hospital volume on ovarian cancer survival and utilized the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database to review of 2952 patients aged 65 and older undergoing surgery for primary ovarian cancer. After adjusting for case complexity and other variables, hospital case volume was not significantly associated with 2-year mortality or overall survival outcome. In contrast to the current study, the analysis from Schrag et al. did not account for patients receiving a treatment paradigm other than up-front surgery, which may have artificially elevated the survival results for those institutions selecting predominantly healthy patients with more limited disease burdens for the primary surgery treatment paradigm. Another potential limitation of the Schrag et al. study is that the analysis controlled for many of the treatment-related variables possibly intrinsic to highvolume centers, such as variations in chemotherapy use and individual surgeon case volume, which would be expected to result in a more favorable survival outcome. In other words, once all of the potentially advantageous characteristics of high-volume centers are removed from the analysis or controlled for, it is not surprising that high-volume center status had minimal impact on survival. The current analysis of 45,929 patients with advanced-stage epithelial ovarian cancer is one of the largest population-based studies of prospectively collected ovarian cancer data reported to date. Our findings indicate that increasing hospital ovarian cancer procedure volume is an independent predictor of improved survival outcome after controlling for the effects of treatment paradigm as well as other important demographic and tumor-related prognostic factors. A secondary objective of the current study was to correlate hospital procedure volume with the likelihood of receiving standard recommended care according to the National Comprehensive Cancer Network (NCCN) ovarian cancer guidelines [30]. Within this framework, the preferred treatment paradigm for patients with FIGO/AJCC Stage IIIC/IV disease is initial surgery with hysterectomy, bilateral salpingo-oophorectomy, surgical staging, and attempted cytoreduction followed by adjuvant platinum/taxane chemotherapy. For patients thought to have primarily unresectable disease, administration of up-front chemotherapy (neo-adjuvant chemotherapy) is considered an alternative treatment strategy. The current data indicate that low/intermediate volume hospitals were significantly more likely to employ a treatment paradigm other than the recommended initial surgery followed by adjuvant chemotherapy. The observed associations between hospital procedure volume and ethnicity and median household income are thought-provoking and require further study to determine their clinical significance. There are several limitations of the current study that must be considered in interpreting the data presented. First, although the NCDB captures approximately 80% of newly diagnosed cancer case in the United States, we cannot discount the possibility of selection bias in submission of individual cases, although this seems unlikely. Similarly, the ovarian cancer data set, like all NCDB data sets, contains incomplete information for some demographic and clinical-pathologic characteristics, which may have introduced another layer of potential
section bias in calculating associations between these factors and hospital procedure volume. A third limitation is that the NCDB does not provide information on individual surgeon case volume or surgical specialty, and we were not able to account for the potential impact of these characteristics on survival outcome or compliance with recommended process measures. Finally, a fourth limitation is that the NCDB does not contain information on the extent of residual disease following attempted cytoreductive surgery. Given the association between increasing hospital procedure volume and a higher likelihood of successful surgical resection to minimal residual disease documented by other investigators, it is possible that differences along this unmeasured parameter may account, at least in part, for the observed positive volume–outcome relationship. Despite these limitations, the current analysis offers a unique perspective on the potential benefit associated with concentration of ovarian cancer care services. Specifically, hospital surgical procedure volume ≥21 cases/year is associated with a higher likelihood of patients with Stage IIIC/IV epithelial ovarian cancer receiving standard treatment (surgery followed by adjuvant chemotherapy). Furthermore, even after adjusting for treatment paradigm and other factors, hospital volume ≥21 cases/year was significantly predictive of improved overall survival outcome. Unlike the investment required for breakthrough anti-neoplastic therapies, concentration of ovarian cancer care requires a commitment to restructuring health service models to optimize patient outcomes. For example, Kommoss et al. [31] have shown that implementation of an institutional quality assurance and disease management program for ovarian cancer care is feasible and associated with a significant improvement in the degree of surgeon and institutional compliance with recommended treatment guidelines. The observation in the current study that only 61.6% of all patients were documented to have received the NCCN recommended treatment paradigm is concerning. While this apparent deviation may reflect some component of incomplete data capture (the surgery and chemotherapy sequence was unknown for 11.4% of patients), it nevertheless highlights the need for overall improvement of cancer care delivery, in addition to concentration of services. The feasibility of centralizing primary surgical care for ovarian cancer on a regional scale with resultant improvements in patient care and survival outcome has been elegantly demonstrated by Tingulstad and colleagues [12,32]. The reality is that hospital procedure volume, like surgeon volume and even surgeon specialty, is still too imprecise a measure of true health care quality. Ultimately, more specific measures of clinically meaningful outcomes coupled with documentation of the associated cost of effective care will need to be developed and rigorously validated. For ovarian cancer, this has been, and will continue to be, an iterative process requiring labor-intensive, prospective, multi-institutional data collection and analysis. Until more robust data are available, concentration of services within high and very high volume centers offers the real possibility of modest but meaningful improvement in survival outcomes for women with ovarian cancer. Conflict of interest statement No authors have a conflict of interest to declare.
References [1] Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics 2009. CA Cancer J Clin 2008;59:225–49. [2] Hillner BE, Smith TJ, Desch CE. Hospital and physician volume of specialization and outcomes in cancer treatment: importance in quality of cancer care. J Clin Oncol 2000;18:2327–40. [3] Carney ME, Lancaster JM, Ford C, Tsodikov A, Wiggins CL. A population-based study of patterns of care for ovarian cancer: who is seen by a gynecologic oncologist and who is not? Gynecol Oncol 2002;84:36–42. [4] du Bois A, Rochon J, Pfisterer J, Hoskins WJ. Variations in institutional infrastructure, physician specialization and experience, and outcome in ovarian cancer: a systematic review. Gynecol Oncol 2009;112:422–36. [5] Kumpulainen S, Grenman S, Kyyronen P, Pukkula E, Sankila R. Evidence of benefit from centralized treatment of ovarian cancer: a nationwide population-based survival analysis in Finland. Int J Cancer 2002;102:541–4.
R.E. Bristow et al. / Gynecologic Oncology 118 (2010) 262–267 [6] Oberainger W, Stuhlinger W. Influence of department volume on cancer survival for gynaecological cancers—a population-based study in Tyeol, Austria. Gynecol Oncol 2006;103:527–34. [7] Ioka A, Tsukuma H, Ajiki W, Oshima A. Influence of hospital procedure volume on ovarian cancer survival in Japan, a country with a low incidence of ovarian cancer. Cancer Sci 2004;95:233–7. [8] Goff BA, Matthews BJ, Wynn M, Muntz HG, Lishner DM, Baldwin LM. Ovarian cancer: patterns of surgical care across the United States. Gynecol Oncol 2006;103:383–90. [9] Goff BA, Matthews BJ, Larson EH, et al. Predictors of comprehensive surgical treatment in patients with ovarian cancer. Cancer 2007;109:2031–42. [10] Marth C, Hieble S, Oberainger W, Winter R, Leodolter S, Sevelda P. Influence of department volume on survival for ovarian cancer: results from a prospective quality assurance program of the Austrian Association for Gynecologic Oncology. Int J Gynecol Cancer 2009;19:94–102. [11] Paulsen T, Kjaerheim K, Kaern J, Tretli S, Trope C. Improved short-term survival for advanced ovarian, tubal, and peritoneal cancer patients operated at teaching hospitals. Int J Gynecol Cancer 2006;16:11–7. [12] Tingulstad S, Skjeldestad FE, Hagen B. The effect of centralization of primary surgery on survival in ovarian cancer patients. Obstet Gynecol 2003;102:499–505. [13] Schrag D, Earle C, Xu F, et al. Associations between hospital and surgeon procedure volumes and patient outcomes after ovarian cancer resection. J Natl Cancer Inst 2006;98:163–71. [14] Vernooj F, Witteveen PO, Verweij E, van der Graaf Y, Heintz APM. The impact of hospital type on the efficacy of chemotherapy treatment in ovarian cancer patients. Gynecol Oncol 2009;115:343–8. [15] Kumpulainen S, Sankila R, Leminen A, et al. The effect of hospital operative volume, residual tumor and first-line chemotherapy on survival of ovarian cancer—a prospective nation-wide study in Finland. Gynecol Oncol 2009;115:199–203. [16] Partridge EE, Phillips JL, Menck HR. The National Cancer Data Base report on ovarian cancer treatment in United States hospitals. Cancer 1996;78:2236–46. [17] Averette HE, Janicek MF, Menck HR. The National Cancer Data Base report on ovarian cancer. Cancer 1995;76:1096–103. [18] Hightower RD, Nguyen HN, Averette HE, Hoskins W, Harrison T, Steren A. National survey of ovarian carcinoma IV: patterns of care and related survival for older patients. Cancer 1994;73:377–83. [19] Parham G, Phillips JL, Hicks ML, et al. The National Cancer Data Base report on malignant epithelial ovarian carcinoma in African-American women. Cancer 1997;80:816–26.
267
[20] Rodriguez M, Nguyen HN, Averette HE, et al. National survey of ovarian carcinoma XII. Cancer 1994;73:1245–50. [21] Cox DR. Regression models and life-tables (with discussion). J Roy Stat Soc (B) 1972;34:187–220. [22] Hoskins W, Rice L, Rubin S. Ovarian cancer surgical practice guidelines. Oncology 1997;11:896–904. [23] Junor EJ, Hole DJ, Gillis CR. Management of ovarian cancer: referral to a multidisciplinary team matters. Br J Cancer 1994;70:363–70. [24] Eisenkop SM, Spirtos NM, Montag TW, Nalick RH, Wang HJ. The impact of subspecialty training on the management of advanced ovarian cancer. Gynecol Oncol 1992;47:203–9. [25] Junor EJ, Hole DJ, McNulty L, Mason M, Young J. Specialist gynaecologists and survival outcome in ovarian cancer: a Scottish national study of 1866 patients. Br J Obstet Gynaecol 1999;106:1130–6. [26] Engelen MJ, Kos HE, Willemse PHB, et al. Surgery by consultant gynecologic oncologists improves survival in patients with ovarian carcinoma. Cancer 2006;106:589–98. [27] Wimberger P, Lehmann N, Kimmig R, Burges A, Meier W, DuBois A. Prognostic factors for complete debulking in advanced ovarian cancer and it impact on survival. An exploratory analysis of a prospectively randomized phase III study of the Arbeitsgemeinschaft Gynaekologische Onkologie Ovarian Cancer Study Group (AGO-OVAR). Gynecol Oncol 2007;106:69–74. [28] duBois A, Rochon J, Lamparter C, Pfisterer J. Pattern of care and impact of participation in clinical studies on the outcome in ovarian cancer. Int J Gynecol Cancer 2005;15:183–91. [29] Marx C, Bendixen A, Hogdall C, Ottosen C, Kehlet H, Ottosen B. Organisation and quality of primary surgical intervention for ovarian cancer in Denmark. Acta Obstetrica et Gynecologica 2007;86:1496–502. [30] National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology: Ovarian Cancer V.2.2009. www.nccn.org. [31] Kommoss S, Harter P, Traut A, et al. Compliance to consensus recommendations, surgeon's experience, and introduction of a quality assurance and manangement program. Influence on therapy of early-stage ovarian carcinoma. Int J Gynecol Cancer 2009;19:787–93. [32] Tinglustad S, Hagen B, Skjeldestad FE, Halvorsen T, Nustad K, Onsrud M. The riskof-malignancy index to evaluate potential ovarian cancers in local hospitals. Obstet Gynecol 1999;93:448–52.
Contents lists available at ScienceDirect
Gynecologic Oncology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y g y n o
The National Cancer Database report on advanced-stage epithelial ovarian cancer: Impact of hospital surgical case volume on overall survival and surgical treatment paradigm Robert E. Bristow a,⁎, Bryan E. Palis b, Dennis S. Chi c, William A. Cliby d a The Kelly Gynecologic Oncology Service, Departments of Gynecology and Obstetrics and Oncology, The Sidney Kimmel Comprehensive Cancer Center, The Johns Hopkins Medical Institutions, Baltimore, MD, USA b National Cancer Data Base, Commission on Cancer at the American College of Surgeons, Chicago, IL, USA c Gynecology Service Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY, USA d Department of Gynecologic Surgery, Mayo Clinic, Rochester, MN, USA
a r t i c l e
i n f o
Article history: Received 23 April 2010 Available online 22 June 2010 Keywords: Ovarian cancer Surgical volume Survival
a b s t r a c t Objective. To examine the effect of hospital procedure volume and other prognostic variables on overall survival outcome and likelihood of receiving standard recommended care among patients with advancedstage epithelial ovarian cancer. Methods. The National Cancer Data Base (NCDB) was searched for patients undergoing primary treatment for FIGO Stage IIIC/IV epithelial ovarian cancer from 1996 to 2005. The average annual surgical procedure volume was derived for each reporting hospital. Quartile ranking discriminated four groups of hospitals based on annual surgical volume: low (b9), intermediate (9–20), high (21–35), and very high (N 35). Cox proportional hazards modeling was used to determine the impact on overall survival of hospital surgical volume adjusted for treatment, FIGO/AJCC stage, ethnicity, age, payer status, household income, and tumor grade. Binomial multivariate logistic regression modeling was used to assess differences in patient demographic, tumor, and treatment variables between high/very high volume hospitals and low/ intermediate volume hospitals. Results. A total of 45,929 patients were identified. After adjusting for other factors, overall survival was significantly correlated with hospital case volume: very high (reference); high (HR 0.98, 95% CI = 0.92–1.04); intermediate (HR 1.08, 95% CI = 1.01–1.15); and low (HR 1.14, 95% CI = 1.07–1.22). Compared to low and intermediate volume hospitals, patients treated at very high and high-volume hospitals were less likely to receive neo-adjuvant chemotherapy (OR = 0.33, 95% CI = 1.18–1.50) or surgery alone (OR = 0.77, 95% CI = 0.73–0.82) instead of initial surgery and adjuvant chemotherapy. Conclusions. Hospital ovarian cancer surgical volume ≥ 21 cases/year is associated with a higher likelihood of patients with Stage IIIC/IV epithelial ovarian cancer receiving standard treatment (surgery followed by adjuvant chemotherapy). Even after adjusting for treatment paradigm and other factors, hospital volume ≥ 21 cases/year was significantly predictive of improved overall survival outcome. © 2010 Elsevier Inc. All rights reserved.
Introduction The American Cancer Society estimated that 21,550 women in the United States will be diagnosed with ovarian cancer in 2009, and 14,600 women will die of this disease [1]. While advances in novel chemotherapeutic agents and treatment strategies continue to show incremental benefits in survival, recent attention has focused on the potential for improved healthcare outcomes through concentration of
⁎ Corresponding author. Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of California, Irvine - Medical Center, 101 The City Drive, Building 56, Room 260 Orange, CA 92686, USA. Fax: +1 714 456 7754. E-mail address: [email protected] (R.E. Bristow). 0090-8258/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2010.05.025
cancer services. For example, Hillner et al. [2] reported findings of a comprehensive review of the health services literature and found that across all studies, the absolute benefit from care at high-volume centers exceeded the benefit from breakthrough treatments and concluded that efforts to concentrate the initial care for all forms of cancer should be undertaken. A number of population-based studies have demonstrated that the surgical expertise and multidisciplinary care provided to patients with ovarian cancer by high-volume surgeons and high-volume centers is superior when compared to low-volume providers [3–12]. However, other investigators have reported contradictory findings [13–15]. With recent attention focused on improving the quality of cancer care, correlative clinical outcome data for healthcare delivery system characteristics can play an important role in quality improvement efforts. Therefore, the
R.E. Bristow et al. / Gynecologic Oncology 118 (2010) 262–267
potential relationship between hospital ovarian cancer case volume and meaningful clinical outcomes must be clearly defined. The National Cancer Data Base (NCDB) is a joint program of the American College of Surgeons Commission on Cancer and the American Cancer Society and serves to collect data for research and quality improvement. Established in 1985, the NCDB contains data from approximately 20 million patients from more than 1430 participating hospitals, and captures nearly 80% of newly diagnosed cancers in the United States each year. The NCDB collects patient demographics, tumor characteristics, clinical and pathologic TNM stage classification, treatments administered, and survival data [16–20]. The objective of this study was to examine and attempt to quantify the effect of hospital procedure volume on overall survival and processes of care for women with advanced-stage invasive epithelial ovarian cancer within the context of demographic and disease-related prognostic factors using the NCDB. Methods Approval to conduct this study was obtained from the Johns Hopkins Medical Institutions Clinical Research Committee and Joint Commission on Clinical Investigation, and the requirement for informed patient consent was waived (IRB study #NA00037503). Invasive epithelial ovarian patients diagnosed between January 1, 1996, and December 31, 2005, were selected from the NCDB by topography code C56.9 (epithelial histologic types 8010 to 8569; 8940 to 8949; 9000) according to the International Classification of Disease— Oncology 3rd edition (ICD-O-3). Subjects were included if they had an invasive epithelial ovarian malignancy alone or as the first of two or more primary cancers. Patients with borderline ovarian tumors were not included. The study population was further selected to include only those patients with International Federation of Gynecology and Obstetrics (FIGO) and American Joint Committee on Cancer (AJCC) Stage IIIC/IV disease who were managed according to one of the following treatment paradigms: initial surgery followed by adjuvant chemotherapy, initial (neo-adjuvant) chemotherapy followed by surgery, or surgery alone. Surgery was defined as any of the following procedures: local tumor destruction, total removal of a tumor, unilateral oophorectomy, bilateral oophorectomy with omentectomy, debulking/ cytoreductive surgery, pelvic exenteration, or other oophorectomy/surgery type not specified. According to NCDB protocols, hospital cancer registries abstracted cases with the Registry Operations and Data Standards (ROADS) and the Facility Oncology Registry Data Standards (FORDS) manuals, a standardized set of data elements and definitions, using information provided by both patients and hospital medical information systems. Patients received all or part of their first course treatment at the reporting facility. All patients were diagnosed and treated at either community, community comprehensive, or teaching/research hospitals, excluding affiliate and free-standing clinics, pediatric hospitals, and Veterans Affairs facilities. Hospitals were required to have reported data for N7 years between 1996 and 2005. The average annual hospital surgical procedure volume was derived for each reporting hospital based on the total number of patients undergoing operative intervention for all stages of invasive epithelial ovarian cancer and the number of years with reported data. Four groups of surgical volume were identified based on quartile rank; N75th percentile = very high volume (N35 cases/year), 51st– 75th percentile = high volume (21–35 cases/year), 25th–50th percentile = intermediate volume (9–20 cases/year), and b25th percentile = low volume (b9 cases/year). Median household income estimates of patient zip code of residence at the time of diagnosis were obtained by linking the NCDB to US Census data, as patient income is not captured in the NCDB. Five-year overall survival rates were calculated from date of diagnosis to date of death or last follow-up. Patients diagnosed from
263
1996 to 2000 were evaluated for survival outcome, as this group of patients had at minimum 5 years of follow-up data. Survival was estimated by the life table method, and statistical comparisons of survival between hospital volume groups were performed with logrank pairwise comparisons. Cox proportional hazards modeling was used to determine the risk of mortality according to hospital surgical volume adjusted for treatment, pathologic stage, ethnicity, age, payer status, household income, and tumor grade [21]. Hazard ratios (HR) with 95% confidence intervals (95% CI) were generated; HR N 1.0 indicates an increased likelihood of death. Binomial Logistic regression modeling was used to assess patient demographic and tumor characteristics associated with hospital procedure volume categories, adjusted for treatment paradigm, pathologic stage, ethnicity, age, payer status, household income, and tumor grade. Odds ratios (OR) with 95% CIs were generated; OR N 1.0 indicates an increased likelihood of chance. The level of statistical significance was set at p b 0.05. All analyses were generated using SPSS 17.0.1 (SPSS Inc, Chicago, IL). Results At the time of analysis, the NCDB contained 111,557 adult female patients aged ≥18 years at the time of invasive epithelial ovarian cancer diagnosis from 1996 through 2005. Of that total patient pool, 45,929 patients were classified as having pathologic FIGO/AJCC Stage IIIC or IV disease and were treated with either surgery followed by chemotherapy, neo-adjuvant chemotherapy followed by surgery, or surgery alone and were treated at hospitals with ≥7 years of reported data. A total of 1184 hospitals reported sufficient data during the study interval: very high volume, n = 76 (6.4%); high volume, n = 148 (12.5%); intermediate volume, n = 304 (25.7%); and low volume, n = 656 (55.1%). Hospital classification, patient demographics, and tumor characteristics are shown in Table 1. Survival analysis A total of 20,600 patients were available for long-term survival analysis from the 1996–2000 cohort with a minimum of 5-years of follow-up. Log-rank pairwise comparisons revealed that 5-year survival rates for patients treated at low and intermediate volume hospitals (b21 surgeries annually) experienced poorer survival outcomes than patients from high and very high volume hospitals (p b 0.001) (Table 2, Fig. 1). Although very high volume hospitals (N35 surgeries annually) had superior overall survival compared to highvolume hospitals (21–35 surgeries annually), this difference only approached statistical significance (p = 0.07). The 1996–2000 patient cohort was examined by a multivariate Cox regression analysis to determine significant and independent associations between overall survival and demographic and clinicalpathologic variables, including hospital procedure volume. After removing patients with missing values, 12,276 patients were available for analysis. Treatment paradigm, pathologic FIGO/AJCC stage, ethnicity, age at diagnosis, primary payer, household income level, and tumor grade were significantly associated with overall survival (Table 3). The survival outcome of patients treated with either neoadjuvant chemotherapy or surgery alone was significantly worse compared to the treatment paradigm of initial surgery followed by adjuvant chemotherapy. After adjusting for the effects of other factors, there was a statistically significant and independent association between increasing hospital procedure volume and overall survival. Using very high volume hospitals as the referent group, there was no significant difference in overall survival compared to high-volume hospitals. However, below a hospital surgical volume of 21 cases/year, there was a stepwise decrease in overall survival as the hospital procedure volume decreased: intermediate volume HR = 1.08 (95% CI = 1.01 to 1.15), low-volume HR = 1.14 (95% CI = 1.07 to 1.22).
264
R.E. Bristow et al. / Gynecologic Oncology 118 (2010) 262–267
Table 1 Hospital volume, patient demographics, and tumor characteristics for epithelial ovarian cancer Stage IIIC/IV, 1996–2005 (n = 45,929).
Treatment paradigm
AJCC pathologic stage Tumor grade
Ethnicity
Age at diagnosis Primary payer
Household income
Surgery only Neoadjuvant chemotherapy + surgery Surgery + adjuvant chemotherapy Surgery–chemotherapy sequence unknown Stage IIIC Stage IV Well/moderately differentiated Undifferentiated/anaplastic Unknown Non-Hispanic White Non-Hispanic Black Non-Hispanic other Hispanic Unknown Mean (years) Private insurance Managed care Medicaid Medicare/Medicare with supplement Not Insured Unknown N$35,000/year b$34,999/year Unknown
Total
Very high
High
Intermediate
Low
Total
2416 (18.1%) 867 (6.5%) 8365 (62.5%) 1737 (13.0%) 9482 (70.8%) 3903 (29.2%) 2606 (19.5%) 8876 (66.3%) 1903 (14.2%) 10774 (80.5%) 730 (5.5%) 211 (1.6%) 452 (3.4%) 1218 (9.1%) 61.6 2497 (18.7%) 4172 (31.2%) 525 (3.9%) 4653 (34.8%) 468 (3.5%) 1070 (8.0%) 3787 (28.3%) 8755 (65.4%) 843 (6.3%) 13385
3048 (21.8%) 745 (5.3%) 8963 (64.2%) 1203 (8.6%) 9634 (69.0%) 4325 (31.0%) 2913 (20.9%) 9013 (64.6%) 2033 (14.6%) 10916 (78.2%) 764 (5.5%) 330 (2.4%) 518 (3.7%) 1431 (10.3%) 62.2 2668 (19.1%) 4355 (31.2%) 510 (3.7%) 5357 (38.4%) 397 (2.8%) 672 (4.8%) 3733 (26.7%) 9376 (67.2%) 850 (6.1%) 13959
2717 (23.9%) 573 (5.1%) 6602 (58.2%) 1454 (12.8%) 7512 (66.2%) 3834 (33.8%) 2372 (20.9%) 7038 (62.0%) 1936 (17.1%) 8828 (77.8%) 738 (6.5%) 276 (2.4%) 607 (5.3%) 897 (7.9%) 62.7 2044 (18.0%) 3425 (30.2%) 469 (4.1%) 4355 (38.4%) 523 (4.6%) 530 (4.7%) 3084 (27.2%) 7640 (67.3%) 622 (5.5%) 11346
1722 (23.8%) 332 (4.6%) 4360 (60.2%) 825 (11.4%) 4114 (56.8%) 3125 (43.2%) 1547 (21.4%) 4116 (56.9%) 1576 (21.8%) 5778 (79.8%) 391 (5.4%) 149 (2.1%) 275 (3.8%) 646 (8.9%) 64 1344 (18.6%) 1661 (22.9%) 334 (4.6%) 3258 (45.0%) 244 (3.4%) 398 (5.5%) 1486 (20.5%) 5379 (74.3%) 374 (5.2%) 7239
9903 2517 28290 5219 30742 15187 9438 29043 7448 36296 2623 966 1852 4192
Characteristics associated with hospital procedure volume The entire ovarian cancer study population was examined for significant independent associations between hospital procedure volume and demographic and clinical-pathologic variables. After excluding patients with missing values, a total of 27,161 patients were available for study. For the purposes of this analysis, very high and high-volume hospitals were combined into a single category, and intermediate and low-volume hospitals were combined into a single category (Table 4). Compared to patients treated at low/intermediate volume hospitals, patients treated at very high/high-volume hospitals were less likely to be treated with neo-adjuvant chemotherapy (OR = 0.33, 95% CI = 0.18 to 0.50) and with surgery alone (OR = 0.77,
(21.6%) (5.5%) (61.6%) (11.4%) (66.9%) (33.1%) (20.5%) (63.2%) (16.25) (79.0%) (5.7%) (2.1%) (4.0%) (9.1%)
8553 (18.6%) 13613 (29.6%) 1838 (4.0%) 17623 (38.4%) 1632 (3.6%) 2670 (5.8%) 12090 (26.3%) 31150 (67.8%) 2689 (5.9%) 45929
95% CI = 0.73 to 0.82) versus surgery followed by adjuvant chemotherapy. In addition, patients from very high/high surgical volume hospitals were less likely to have pathologic stage IV disease and undifferentiated/anaplastic tumors, be of Hispanic ethnicity, be age N 64 years, and have a lower median household income level. Discussion Concentration of cancer care services has been advocated on the basis of improved clinical outcomes associated with clinicians and centers providing high-volume clinical practices. With regard to ovarian cancer, the Society of Surgical Oncology has provided the following guidelines for surgery: “Surgeons undertaking operations
Table 2 Log-rank pairwise comparisons of observed overall survival rates for epithelial ovarian cancer Stage IIIC/IV, 1996–2000 (n = 17,683).
Very high volume
High volume
Intermediate volume
Low volume
Month of survival
Percent survival rate (95% CI)
0 12 24 36 48 60 0 12 24 36 48 60 0 12 24 36 48 60 0 12 24 36 48 60
100% 82.1% 64.1% 48.8% 37.2% 28.9% 100% 79.3% 61.7% 46.6% 35.9% 28.5% 100% 78.6% 58.4% 43.3% 32.9% 26.1% 100% 75.0% 55.6% 40.8% 31.3% 24.3%
Very high volume
High volume
Intermediate volume
Low volume
(81.2%–83.0%) (63.0%–65.3%) (47.6%–50.0%) (36.0%–38.4%) (27.8%–30.1%)
NA
0.07
b0.001
b 0.001
(78.2%–80.5%) (60.4%–63.1%) (45.2%–48.0%) (34.6%–37.3%) (27.2%–29.8%)
0.07
NA
b0.001
b 0.001
(77.1%–80.0%) (56.7%–60.2%) (41.6%–45.1%) (31.3%–34.6%) (24.5%–27.7%)
0.00
b0.001
NA
0.02
(73.3%–76.7%) (53.7%–57.5%) (38.9%–42.8%) (29.5%–33.1%) (22.6%–26.0%)
b 0.001
b0.001
0.02
NA
R.E. Bristow et al. / Gynecologic Oncology 118 (2010) 262–267
265
Table 4 Comparison (binomial logistic regression model) of intermediate/low-volume hospitals (baseline) and very high/high-volume hospitals according to patient demographics and tumor characteristics for epithelial ovarian cancer Stage IIIC/IV, 1996–2005 (n = 27,161). Characteristic
Fig. 1. Overall survival according to hospital surgical case volume for epithelial ovarian cancer Stage IIIC/IV, 1996–2000 (n = 20,600).
for possible ovarian cancer should have both the necessary technical expertise and a thorough understanding of the management of the disease itself…optimal treatment of this disease requires the skillful and appropriate integration of cancer surgery and chemotherapy, and is best carried out in centers in which an experienced and coordinated multidisciplinary team is available” [22]. The association between subspecialist care by a gynecologic oncologist and superior ovarian cancer outcomes has been well documented [3,23–26]. Recently,
Table 3 Risk factors for mortality for epithelial ovarian cancer Stage IIIC/IV, 1996–2000 (n = 10,641). Risk factor Hospital surgical case volume Very high High Intermediate Low Treatment paradigm Surgery + adjuvant chemotherapy Neoadjuvant chemotherapy + surgery Surgery alone FIGO/AJCC stage group Stage IIIC Stage IV Tumor grade Well/moderately differentiated Undifferentiated/anaplastic Ethnicity Non-Hispanic White Non-Hispanic Black Hispanic Age at diagnosis b40 years 40–64 years N64 years Primary payer Private insurance Managed care Medicaid Medicare/Medicare with supplement Not insured Household income N$35,000/year b$34,999/year
N
Significance level
Hazard ratio (95% CI)
4046 3066 1936 1593
0.26 0.03 0.00
Referent 1.03 (0.98–1.09) 1.08 (1.01–1.15) 1.16 (1.09–1.24)
5570 300 4771
0.00 0.01
Referent 1.20 (1.15–1.26) 1.19 (1.04–1.36)
6994 3647
0.00
Referent 1.42 (1.35–1.48)
2871 7770
0.00
Referent 1.16 (1.10–1.22)
9584 670 387
0.01 0.09
Referent 1.14 (1.04–1.25) 0.89 (0.79–1.02)
383 5141 5117
0.21 0.00
Referent 1.09 (0.95–1.25) 1.43 (1.23–1.66)
2465 2935 4505 353 383
0.90 0.00 0.00 0.01
Referent 1.00 (0.93–1.07) 1.35 (1.19–1.54) 1.15 (1.06–1.26) 1.19 (1.03–1.37)
3030 7611
0.00
Referent 1.09 (1.03–1.15)
Treatment paradigm Surgery + adjuvant chemotherapy Neoadjuvant chemotherapy + surgery Surgery alone FIGO/AJCC stage group Stage IIIC Stage IV Tumor grade Well/moderately differentiated Undifferentiated/anaplastic Ethnicity Non-Hispanic White Non-Hispanic Black Hispanic Age at diagnosis b40 years 40–64 years N64 years Primary payer Private insurance Managed care Medicaid Medicare/Medicare with supplement Not insured Household income N$35,000/year b$34,999/year
N
Significance level
Odds ratio (95% CI)
19213 1473 6475
0.00 0.00
Referent 0.33 (0.18–0.50) 0.77 (0.73–0.82)
18369 8792
0.00
Referent 0.72 (0.69–0.76)
6699 20462
0.00
Referent 0.86 (0.81–0.91)
24292 1744 1125
0.12 0.00
Referent 0.92 (0.83–1.02) 0.81 (0.72–0.92)
883 13685 12593
0.19 0.00
Referent 0.91 (0.78–1.05) 0.78 (0.66–0.92)
5454 8399 1205 11002 1101
0.00 0.64 0.37 0.00
Referent 1.14 (1.06–1.23) 0.97 (0.84–1.11) 1.05 (0.95–1.15) 0.80 (0.70–0.93)
7594 19567
0.00
Referent 0.89 (0.84–0.95)
however, the medical community focused on the potential impact of hospital concentration of ovarian cancer care on patient outcomes. Despite this focus, it has been difficult to quantify the possible benefit from hospital concentration of cancer services as a result of the multifactorial survival determinants for patients with ovarian cancer. Given the complexity of surgical and post-surgical care required for the effective management of patients with advanced-stage ovarian cancer in particular, the primary objective of the current study was to explore a potential positive hospital volume–outcome relationship for this group of patients using the resources of the NCDB. The majority of published population-based studies evaluating hospital procedure volume and ovarian cancer outcomes have come from outside of the United States, and the results have been somewhat contradictory [4–7,10–13,27–29]. For example, Marth et al. [10] reported a prospective study of 1948 patients in Austria. After controlling for other prognostic variables, treatment at a low-volume center (b24 cases/year) was a statistically significant and independent negative prognostic factor for overall survival (HR = 1.38, 95% CI = 1.2–1.7, p b 0.001). Ioka et al. [7] reported on 2450 ovarian cancer patients from the Osaka Cancer Registry. These investigators found that, after controlling for other variables, patients cared for at very low-volume hospitals (b1 case/year) had a 60% increase in the risk of disease-related death compared to high-volume hospitals. Similarly, Wimberger et al. [27] reported on 785 patients from Germany, extracted from the AGO-OVAR Study Group, and found that surgery at a high-volume center was associated with a significant reduction in the risk of disease-related death (HR = 0.77, 95% CI = 0.63–0.94, p = 0.012) after controlling for other variables. Not all studies have found a significant correlation between hospital volume and ovarian cancer survival outcome, however. For example, Vernooj et al. [14] investigated the influence of hospital case volume on ovarian cancer survival among 1077 patients from 18 Dutch hospitals between 1996 and 2003. High-volume hospital (≥13 cases/year) care was associated with a higher likelihood of adequate
266
R.E. Bristow et al. / Gynecologic Oncology 118 (2010) 262–267
staging procedures but was not significantly associated with overall survival. Similarly, Kumpulainen et al. [15] conducted a follow-up study of 275 ovarian cancer patients from the Finish Cancer Registry and found that an annual hospital ovarian cancer operative case volume of ≥10 cases was significantly predictive of complete surgical resection but was not an independent predictor of overall survival. There are two U.S. population-based ovarian cancer studies examining the potential impact of hospital case volume [8,13]. Goff et al. [8] reported an analysis of 10,432 patients from the Health Care Cost and Utilization Project investigating the impact of hospital case volume on stage-appropriate surgical procedures. These authors found that, after controlling for other factors, patients operated upon at a highvolume hospital were more likely to undergo lymphadenectomy and tumor cytoreduction [8]. In 2006, Schrag et al. [13] published the only large U.S. population-based outcome study to examine the impact of hospital volume on ovarian cancer survival and utilized the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database to review of 2952 patients aged 65 and older undergoing surgery for primary ovarian cancer. After adjusting for case complexity and other variables, hospital case volume was not significantly associated with 2-year mortality or overall survival outcome. In contrast to the current study, the analysis from Schrag et al. did not account for patients receiving a treatment paradigm other than up-front surgery, which may have artificially elevated the survival results for those institutions selecting predominantly healthy patients with more limited disease burdens for the primary surgery treatment paradigm. Another potential limitation of the Schrag et al. study is that the analysis controlled for many of the treatment-related variables possibly intrinsic to highvolume centers, such as variations in chemotherapy use and individual surgeon case volume, which would be expected to result in a more favorable survival outcome. In other words, once all of the potentially advantageous characteristics of high-volume centers are removed from the analysis or controlled for, it is not surprising that high-volume center status had minimal impact on survival. The current analysis of 45,929 patients with advanced-stage epithelial ovarian cancer is one of the largest population-based studies of prospectively collected ovarian cancer data reported to date. Our findings indicate that increasing hospital ovarian cancer procedure volume is an independent predictor of improved survival outcome after controlling for the effects of treatment paradigm as well as other important demographic and tumor-related prognostic factors. A secondary objective of the current study was to correlate hospital procedure volume with the likelihood of receiving standard recommended care according to the National Comprehensive Cancer Network (NCCN) ovarian cancer guidelines [30]. Within this framework, the preferred treatment paradigm for patients with FIGO/AJCC Stage IIIC/IV disease is initial surgery with hysterectomy, bilateral salpingo-oophorectomy, surgical staging, and attempted cytoreduction followed by adjuvant platinum/taxane chemotherapy. For patients thought to have primarily unresectable disease, administration of up-front chemotherapy (neo-adjuvant chemotherapy) is considered an alternative treatment strategy. The current data indicate that low/intermediate volume hospitals were significantly more likely to employ a treatment paradigm other than the recommended initial surgery followed by adjuvant chemotherapy. The observed associations between hospital procedure volume and ethnicity and median household income are thought-provoking and require further study to determine their clinical significance. There are several limitations of the current study that must be considered in interpreting the data presented. First, although the NCDB captures approximately 80% of newly diagnosed cancer case in the United States, we cannot discount the possibility of selection bias in submission of individual cases, although this seems unlikely. Similarly, the ovarian cancer data set, like all NCDB data sets, contains incomplete information for some demographic and clinical-pathologic characteristics, which may have introduced another layer of potential
section bias in calculating associations between these factors and hospital procedure volume. A third limitation is that the NCDB does not provide information on individual surgeon case volume or surgical specialty, and we were not able to account for the potential impact of these characteristics on survival outcome or compliance with recommended process measures. Finally, a fourth limitation is that the NCDB does not contain information on the extent of residual disease following attempted cytoreductive surgery. Given the association between increasing hospital procedure volume and a higher likelihood of successful surgical resection to minimal residual disease documented by other investigators, it is possible that differences along this unmeasured parameter may account, at least in part, for the observed positive volume–outcome relationship. Despite these limitations, the current analysis offers a unique perspective on the potential benefit associated with concentration of ovarian cancer care services. Specifically, hospital surgical procedure volume ≥21 cases/year is associated with a higher likelihood of patients with Stage IIIC/IV epithelial ovarian cancer receiving standard treatment (surgery followed by adjuvant chemotherapy). Furthermore, even after adjusting for treatment paradigm and other factors, hospital volume ≥21 cases/year was significantly predictive of improved overall survival outcome. Unlike the investment required for breakthrough anti-neoplastic therapies, concentration of ovarian cancer care requires a commitment to restructuring health service models to optimize patient outcomes. For example, Kommoss et al. [31] have shown that implementation of an institutional quality assurance and disease management program for ovarian cancer care is feasible and associated with a significant improvement in the degree of surgeon and institutional compliance with recommended treatment guidelines. The observation in the current study that only 61.6% of all patients were documented to have received the NCCN recommended treatment paradigm is concerning. While this apparent deviation may reflect some component of incomplete data capture (the surgery and chemotherapy sequence was unknown for 11.4% of patients), it nevertheless highlights the need for overall improvement of cancer care delivery, in addition to concentration of services. The feasibility of centralizing primary surgical care for ovarian cancer on a regional scale with resultant improvements in patient care and survival outcome has been elegantly demonstrated by Tingulstad and colleagues [12,32]. The reality is that hospital procedure volume, like surgeon volume and even surgeon specialty, is still too imprecise a measure of true health care quality. Ultimately, more specific measures of clinically meaningful outcomes coupled with documentation of the associated cost of effective care will need to be developed and rigorously validated. For ovarian cancer, this has been, and will continue to be, an iterative process requiring labor-intensive, prospective, multi-institutional data collection and analysis. Until more robust data are available, concentration of services within high and very high volume centers offers the real possibility of modest but meaningful improvement in survival outcomes for women with ovarian cancer. Conflict of interest statement No authors have a conflict of interest to declare.
References [1] Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics 2009. CA Cancer J Clin 2008;59:225–49. [2] Hillner BE, Smith TJ, Desch CE. Hospital and physician volume of specialization and outcomes in cancer treatment: importance in quality of cancer care. J Clin Oncol 2000;18:2327–40. [3] Carney ME, Lancaster JM, Ford C, Tsodikov A, Wiggins CL. A population-based study of patterns of care for ovarian cancer: who is seen by a gynecologic oncologist and who is not? Gynecol Oncol 2002;84:36–42. [4] du Bois A, Rochon J, Pfisterer J, Hoskins WJ. Variations in institutional infrastructure, physician specialization and experience, and outcome in ovarian cancer: a systematic review. Gynecol Oncol 2009;112:422–36. [5] Kumpulainen S, Grenman S, Kyyronen P, Pukkula E, Sankila R. Evidence of benefit from centralized treatment of ovarian cancer: a nationwide population-based survival analysis in Finland. Int J Cancer 2002;102:541–4.
R.E. Bristow et al. / Gynecologic Oncology 118 (2010) 262–267 [6] Oberainger W, Stuhlinger W. Influence of department volume on cancer survival for gynaecological cancers—a population-based study in Tyeol, Austria. Gynecol Oncol 2006;103:527–34. [7] Ioka A, Tsukuma H, Ajiki W, Oshima A. Influence of hospital procedure volume on ovarian cancer survival in Japan, a country with a low incidence of ovarian cancer. Cancer Sci 2004;95:233–7. [8] Goff BA, Matthews BJ, Wynn M, Muntz HG, Lishner DM, Baldwin LM. Ovarian cancer: patterns of surgical care across the United States. Gynecol Oncol 2006;103:383–90. [9] Goff BA, Matthews BJ, Larson EH, et al. Predictors of comprehensive surgical treatment in patients with ovarian cancer. Cancer 2007;109:2031–42. [10] Marth C, Hieble S, Oberainger W, Winter R, Leodolter S, Sevelda P. Influence of department volume on survival for ovarian cancer: results from a prospective quality assurance program of the Austrian Association for Gynecologic Oncology. Int J Gynecol Cancer 2009;19:94–102. [11] Paulsen T, Kjaerheim K, Kaern J, Tretli S, Trope C. Improved short-term survival for advanced ovarian, tubal, and peritoneal cancer patients operated at teaching hospitals. Int J Gynecol Cancer 2006;16:11–7. [12] Tingulstad S, Skjeldestad FE, Hagen B. The effect of centralization of primary surgery on survival in ovarian cancer patients. Obstet Gynecol 2003;102:499–505. [13] Schrag D, Earle C, Xu F, et al. Associations between hospital and surgeon procedure volumes and patient outcomes after ovarian cancer resection. J Natl Cancer Inst 2006;98:163–71. [14] Vernooj F, Witteveen PO, Verweij E, van der Graaf Y, Heintz APM. The impact of hospital type on the efficacy of chemotherapy treatment in ovarian cancer patients. Gynecol Oncol 2009;115:343–8. [15] Kumpulainen S, Sankila R, Leminen A, et al. The effect of hospital operative volume, residual tumor and first-line chemotherapy on survival of ovarian cancer—a prospective nation-wide study in Finland. Gynecol Oncol 2009;115:199–203. [16] Partridge EE, Phillips JL, Menck HR. The National Cancer Data Base report on ovarian cancer treatment in United States hospitals. Cancer 1996;78:2236–46. [17] Averette HE, Janicek MF, Menck HR. The National Cancer Data Base report on ovarian cancer. Cancer 1995;76:1096–103. [18] Hightower RD, Nguyen HN, Averette HE, Hoskins W, Harrison T, Steren A. National survey of ovarian carcinoma IV: patterns of care and related survival for older patients. Cancer 1994;73:377–83. [19] Parham G, Phillips JL, Hicks ML, et al. The National Cancer Data Base report on malignant epithelial ovarian carcinoma in African-American women. Cancer 1997;80:816–26.
267
[20] Rodriguez M, Nguyen HN, Averette HE, et al. National survey of ovarian carcinoma XII. Cancer 1994;73:1245–50. [21] Cox DR. Regression models and life-tables (with discussion). J Roy Stat Soc (B) 1972;34:187–220. [22] Hoskins W, Rice L, Rubin S. Ovarian cancer surgical practice guidelines. Oncology 1997;11:896–904. [23] Junor EJ, Hole DJ, Gillis CR. Management of ovarian cancer: referral to a multidisciplinary team matters. Br J Cancer 1994;70:363–70. [24] Eisenkop SM, Spirtos NM, Montag TW, Nalick RH, Wang HJ. The impact of subspecialty training on the management of advanced ovarian cancer. Gynecol Oncol 1992;47:203–9. [25] Junor EJ, Hole DJ, McNulty L, Mason M, Young J. Specialist gynaecologists and survival outcome in ovarian cancer: a Scottish national study of 1866 patients. Br J Obstet Gynaecol 1999;106:1130–6. [26] Engelen MJ, Kos HE, Willemse PHB, et al. Surgery by consultant gynecologic oncologists improves survival in patients with ovarian carcinoma. Cancer 2006;106:589–98. [27] Wimberger P, Lehmann N, Kimmig R, Burges A, Meier W, DuBois A. Prognostic factors for complete debulking in advanced ovarian cancer and it impact on survival. An exploratory analysis of a prospectively randomized phase III study of the Arbeitsgemeinschaft Gynaekologische Onkologie Ovarian Cancer Study Group (AGO-OVAR). Gynecol Oncol 2007;106:69–74. [28] duBois A, Rochon J, Lamparter C, Pfisterer J. Pattern of care and impact of participation in clinical studies on the outcome in ovarian cancer. Int J Gynecol Cancer 2005;15:183–91. [29] Marx C, Bendixen A, Hogdall C, Ottosen C, Kehlet H, Ottosen B. Organisation and quality of primary surgical intervention for ovarian cancer in Denmark. Acta Obstetrica et Gynecologica 2007;86:1496–502. [30] National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology: Ovarian Cancer V.2.2009. www.nccn.org. [31] Kommoss S, Harter P, Traut A, et al. Compliance to consensus recommendations, surgeon's experience, and introduction of a quality assurance and manangement program. Influence on therapy of early-stage ovarian carcinoma. Int J Gynecol Cancer 2009;19:787–93. [32] Tinglustad S, Hagen B, Skjeldestad FE, Halvorsen T, Nustad K, Onsrud M. The riskof-malignancy index to evaluate potential ovarian cancers in local hospitals. Obstet Gynecol 1999;93:448–52.