Abstract

Background

We conducted a population-based retrospective cohort study to investigate the influence of hospital volume, delay of surgery, and both together on the long-term survival of postoperative cancer patients.

Methods

Using information from the Korea Central Cancer Registry from 2001 through 2005 and the National Health Insurance claim database, we determined survival for 147 682 patients who underwent definitive surgery for any of six cancers.

Results

Regardless of cancer site, surgical patients in low- to medium-volume hospitals showed significantly worse survival [adjusted hazard ratio (aHR) = 1.36–1.86] than those in high-volume hospitals in multivariable analyses. Among the latter, treatment delays > 1 month were not associated with worse survival for stomach, colon, pancreatic, or lung cancer but were for rectal [aHR = 1.28; 95% confidence interval (CI), 1.17–1.40] and breast (aHR = 1.59; 95% CI, 1.37–1.84) cancer. For patients in low- to medium-volume hospitals, treatment delay was associated with worse survival for all types of cancer (aHR = 1.78–3.81).

Conclusion

Our findings suggest that the effect of hospital volume and surgical treatment delay on overall survival of cancer patients should be considered in formulating or revising national health policy.

introduction

Over the past 30 years, an association between high hospital volume and better outcomes for surgical cancer patients has been demonstrated [1] and reported specifically for oral, breast [2, 3], colon [4–6], esophageal [5, 6], lung [5, 6], liver [5], pancreatic [5–7], rectal [5], bladder [6], and gastric [5, 6] cancer. As a result, the past decade has seen extensive centralization of complex cancer surgery [1], and some countries have implemented policies to achieve this goal [1, 2, 8, 9].

Centralization, however, can substantially increase travel distance for patients [1, 10], particularly those living in rural areas, and might lead to therapeutic delay for many [10, 11]. That, in turn, could lead to treatment delay and thus decrease patient survival by enabling further tumor invasion [12, 13]. Lengthy waits might also cause psychological distress [13]. Therefore, surgical wait times also have become a frequently targeted quality indicator in health care systems [14].

Despite the substantial body of literature investigating the effect of hospital volume or waiting time on cancer patient outcomes, to the best of our knowledge, no study has examined their concurrent effects.

In a population-based retrospective cohort study, we used the Korea Central Cancer Registry (KCCR) data and National Health Insurance (NHI) data to investigate the influence of hospital volume, delay of surgery, and both together on the long-term survival of postoperative cancer patients.

methods

study patients and database

We identified cancer patients who were registered in the KCCR from 2001 through 2005 and were followed up until the end of 2006. The KCCR is a nationwide population-based cancer registry that was first established in 1980 by the Minister of Health and Welfare in Korea. The number of participating hospitals and registered malignancies has been increasing year by year, with > 180 hospitals currently participating in KCCR. It covers at least 95% of the newly diagnosed malignancies in Korea [15]. The KCCR has patient information including age, sex, region of residence, type of cancer, and date of diagnosis. Details of the history, objectives, and activities of the KCCR have been documented elsewhere [16].

We classified the conditions according to the International Classification of Diseases for Oncology, 3rd edition, and then converted them to the International Classification of Diseases 10th edition (ICD-10) [17]. We obtained the NHI database from the Health Insurance Review and Assessment Service and merged both sets of data using a unique patient identifier. We used the database to select patients aged ≥20 years who had been diagnosed with cancer of the stomach (ICD-10 code: C16), colon (C18), rectum (C19, C20), pancreas (C25), lung (C33 and C34), or breast (C50). Those sites were selected as major sites of interest in Korea at a focus meeting of health professionals that included medical oncologists and surgeons. We excluded patients with multiple cancers and patients who did not undergo cancer surgery as their first definitive treatment.

We obtained patient characteristics (age, sex, region of residence, type of cancer, and date of diagnosis) from the KCCR, the date and cause of death from the Korea National Statistical Office database, and information about primary cancer, date of first treatment, insurance, type of hospital, type of treatment, and comorbidities from the Health Insurance Review and Assessment Service database, which covers almost all Koreans. We excluded from this merged database date-error cases and paper-claim cases. To simplify the presentation of our results, we present here the analyses for cancer surgery (± radiotherapy or chemotherapy); we excluded patients who had only radiotherapy, only chemotherapy, or radio and chemotherapy without cancer surgery. We identified 147 682 patients eligible for the study (Figure 1).

Figure 1.

Flow of study population.

Figure 1.

Flow of study population.

We first categorized age group using the cut-off of 65 years as the criteria for the elderly. We then categorized according to the age distribution of the patients: under 50, from 50 to 64, and over 65 years of age. We grouped patients by sex, year of diagnosis, type of cancer, type of medical care institution (private versus public hospital), hospital type (general, university), hospital volume (number of resections per year), insurance type (NHI, Medical Aid), adjuvant treatment (radiotherapy or chemotherapy), and comorbidities (Charlson index) [18].

We defined hospital volume as the mean number of procedures that each hospital carried out on NHI during each of the 5 years. First, we identified all patients who underwent a resection for cancer (stomach, colon, rectum, pancreas, lung, or breast) from 2001 through 2005. Then, we ranked hospitals by year according to the number of each of the procedures carried out on our cohort in order of increasing estimated total volume and defined tertiles of hospital volume (low, medium, and high) by selecting number cut-off points of each the procedures [2, 19, 20]. In sensitivity analysis, we recategorized hospital volume as a binary variable according to the criteria established by the Leapfrog Group for pancreatic resection (11 cases per year [21]) and the potential candidates of volume-based referral for breast (100 [3], 201 [2] per year), lung (25 [6], 67 [19], 84 [5] per year), colon (93 [19], 126 [5], 166 [4] per year), rectal (35 [5] per year), and stomach resection (17 [5, 6] per year).

We defined treatment delay as the time between the date of cancer diagnosis and the date of the initiation of definitive treatment [22]. We based the 31-day standard for treatment delay on the UK's National Health Service cancer waiting time target in the year 2000 [23].

statistical analysis

We used the χ2 test to compare the characteristics of patients who underwent resection in a low-volume hospital with those who underwent resection in a high-volume hospital and of patients who waited > 1 month for treatment with those who waited ≤ 1 month. For multivariate multiple logistic regression, we used categorical indicator factors that showed significant association in univariate analysis. We carried out multivariable Cox proportional hazards modeling to assess the effects of waiting time and hospital volume for each procedure in each treatment year on overall survival, adjusting for age, sex, Charlson score, hospital type, insurance, radiotherapy, chemotherapy, type of medical care institution, year of diagnosis, and waiting time. Five-year survival data are based on cases diagnosed in 2001 only. We used SAS software version 9.1 (SAS Institute Inc, Cary, NC) and SPSS 12.0 software (SPSS Inc., Chicago, IL) for statistical analysis and calculated P values for statistical significance.

results

The mean patient age was 60.0 years (range, 20.0–98.0) for men and 54.4 years (range, 20.0–100.0) for women. The proportion of cancer patients who received surgery in a high-volume hospital changed slightly from 2001 (87.9%) to 2005 (88.7%).

hospital volume and surgical treatment delay

Supplemental Table S1 (available at Annals of Oncology online) shows the results of χ2 analysis of the association of treatment delay and hospital volume with patient characteristics. Patients within low- to medium-volume hospitals were significantly older, had medical aid, had fewer comorbidities, received radiotherapy or chemotherapy, and were admitted to a private hospital. Patients within low- to medium-volume hospitals were more likely to have rectal, lung, or pancreatic cancer and less likely to have stomach or breast cancer. Patients with treatment delays of over 1 month were significantly older, male, received medical aid, had fewer comorbidities, did not receive chemotherapy, and were admitted to a public high-volume hospital. Patients with treatment delays were more likely to have rectal, breast, pancreatic, lung, and stomach cancer than colon cancer. Delays of > 1 month were experienced by 13.6% of the subjects over the study period.

five-year survival according to hospital volume and surgical treatment delay

The overall 5-year survival rate based on cases diagnosed in 2001 was 70.5%. Risk-adjusted long-term survival based on Cox proportional hazards modeling showed worse survival for patients who underwent surgery at low- to medium-volume hospitals than at high-volume hospitals (Table 1).

Table 1.

Five-year survival of surgical cancer patients according to hospital volume and surgical treatment delay

Variable Hospital volumea Surgical treatment delayb 
 HR (95% CI) HR (95% CI) 
Overall   
 5-year survival (%) 70.5  
 Unadjusted HR 1.66 (1.61–1.71) 1.10 (1.07–1.14) 
 Adjustedc 1.38 (1.33–1.43) 1.10 (1.07–1.14) 
Type of cancer   
 Stomach (C16)   
  5-year survival (%) 68.9  
  Unadjusted HR 1.63 (1.55–1.70) 0.94 (0.90–0.98) 
  Adjustedc 1.36 (1.29–1.44) 1.03 (0.99–1.08) 
 Colon (C18)   
  5-year survival (%) 66.0  
  Unadjusted HR 1.47 (1.37–1.57) 1.08 (0.98–1.19) 
  Adjustedc 1.41 (1.29–1.52) 1.10 (1.00–1.21) 
 Rectal (C19, C20)   
  5-year survival (%) 66.3  
  Unadjusted HR 1.52 (1.41–1.63) 1.29 (1.18–1.41) 
  Adjustedc 1.39 (1.27–1.52) 1.28 (1.17–1.40) 
 Pancreatic (C25)   
  5-year survival (%) 16.2  
  Unadjusted HR 1.49 (1.34–1.66) 1.33 (1.16–1.53) 
  Adjustedc 1.26 (1.11–1.43) 1.23 (1.07–1.41) 
 Lung (C33, C34)   
  5-year survival (%) 50.1  
  Unadjusted HR 1.69 (1.56–1.84) 1.12 (1.02–1.23) 
  Adjustedc 1.60 (1.47–1.74) 1.16 (1.06–1.27) 
 Breast (C50)   
  5-year survival (%) 90.5  
  Unadjusted HR 2.12 (1.89–2.39) 1.57 (1.35–1.81) 
  Adjustedc 1.86 (1.62–2.15) 1.59 (1.37–1.84) 
Variable Hospital volumea Surgical treatment delayb 
 HR (95% CI) HR (95% CI) 
Overall   
 5-year survival (%) 70.5  
 Unadjusted HR 1.66 (1.61–1.71) 1.10 (1.07–1.14) 
 Adjustedc 1.38 (1.33–1.43) 1.10 (1.07–1.14) 
Type of cancer   
 Stomach (C16)   
  5-year survival (%) 68.9  
  Unadjusted HR 1.63 (1.55–1.70) 0.94 (0.90–0.98) 
  Adjustedc 1.36 (1.29–1.44) 1.03 (0.99–1.08) 
 Colon (C18)   
  5-year survival (%) 66.0  
  Unadjusted HR 1.47 (1.37–1.57) 1.08 (0.98–1.19) 
  Adjustedc 1.41 (1.29–1.52) 1.10 (1.00–1.21) 
 Rectal (C19, C20)   
  5-year survival (%) 66.3  
  Unadjusted HR 1.52 (1.41–1.63) 1.29 (1.18–1.41) 
  Adjustedc 1.39 (1.27–1.52) 1.28 (1.17–1.40) 
 Pancreatic (C25)   
  5-year survival (%) 16.2  
  Unadjusted HR 1.49 (1.34–1.66) 1.33 (1.16–1.53) 
  Adjustedc 1.26 (1.11–1.43) 1.23 (1.07–1.41) 
 Lung (C33, C34)   
  5-year survival (%) 50.1  
  Unadjusted HR 1.69 (1.56–1.84) 1.12 (1.02–1.23) 
  Adjustedc 1.60 (1.47–1.74) 1.16 (1.06–1.27) 
 Breast (C50)   
  5-year survival (%) 90.5  
  Unadjusted HR 2.12 (1.89–2.39) 1.57 (1.35–1.81) 
  Adjustedc 1.86 (1.62–2.15) 1.59 (1.37–1.84) 

Derived from Cox proportional hazards models for crude (unadjusted) and adjusted analyses.

aLow to medium (versus high). We recategorized hospital volume defined by number of operations/yr with cut-off points (tertiles) of low, medium, and high. The cut-off points are for high-volumes of stomach resection (≥56 per year), colon (≥24 per year), rectal (≥23 per year), pancreatic (≥5 per year), lung (≥15 per year), and breast (≥31 per year).

bSurgical treatment delay > 31 days (versus ≤ 31 days).

cAdjusted for age, sex, Charlson scale, hospital type, insurance, radiotherapy, chemotherapy, type of medical care institution, year of diagnosis, and treatment delay or hospital volume.

CI, confidence interval; HR, hazard ratio.

Among surgery patients, treatment delays > 1 month were associated with worse survival for patients with rectal [adjusted hazard ratio (aHR) = 1.28], pancreatic (aHR = 1.23), lung (aHR = 1.16), or breast (aHR = 1.59) cancer than treatment delays <1 month (Table 1).

survival by hospital volume and surgical treatment delay

Among cancer patients who received surgery in high-volume hospitals, treatment delays > 1 month were associated with worse survival for patients with rectal or breast cancer but not for those with stomach, colon, pancreatic, or lung cancer (Figure 2).

Figure 2.

Survival of patients who had resection of each cancers according to hospital volumea and surgical treatment delayb (adjusted HR, 95% CI). aHR, adjusted hazard ratio; CI, confidence interval. All types of P < 0.01. aWe recategorized hospital volume defined by number of operations per year with cut-off points (tertiles) of low, medium, and high as a binary variable (low–medium versus high). The cut-off points are for high volumes of stomach resection (≥56 per year), colon (≥24 per year), rectal (≥23 per year), pancreatic (≥56 per year), lung (≥15 per year), and breast [≥31 per year]. bSurgical treatment delay > 31 days (versus ≤ 31 days). Adjusted age, sec, Charlson scale, hospital type, insurance, radiotherapy, chemotherapy, type of medical care institutions, year of diagnosis.

Figure 2.

Survival of patients who had resection of each cancers according to hospital volumea and surgical treatment delayb (adjusted HR, 95% CI). aHR, adjusted hazard ratio; CI, confidence interval. All types of P < 0.01. aWe recategorized hospital volume defined by number of operations per year with cut-off points (tertiles) of low, medium, and high as a binary variable (low–medium versus high). The cut-off points are for high volumes of stomach resection (≥56 per year), colon (≥24 per year), rectal (≥23 per year), pancreatic (≥56 per year), lung (≥15 per year), and breast [≥31 per year]. bSurgical treatment delay > 31 days (versus ≤ 31 days). Adjusted age, sec, Charlson scale, hospital type, insurance, radiotherapy, chemotherapy, type of medical care institutions, year of diagnosis.

All cancer patients who received surgery in low- to medium-volume hospitals, regardless of the length of treatment delays, showed worse survival than patients who had treatment delays < 1 month in a high-volume hospital. Patients who had treatment delays > 1 month in low- to medium-volume hospital had the worst survival regardless of cancer site (Figure 2).

sensitivity analysis of overall survival by hospital volume and surgical treatment delay

In sensitivity analysis of the concurrent effects of hospital volume and treatment delay according to recommended volume, overall survival changed only for breast cancer, showing that treatment delay was not associated with worse survival in high-volume hospitals (Table 2).

Table 2.

Sensitivity analysis of hazard ratios for overall survival in patients who had resection for each type of cancer according to surgical treatment delaya and hospital volumeb as a binary variable according to previously established criteria

Variable aHRc (95% CI) aHRc (95% CI) aHRc (95% CI) 
Type of cancer    
Stomach (C16)  ≤ 17   
High and 31 day   
High and > 31 day 1.01 (0.96–1.05)   
Low and ≤ 31 day 1.54 (1.41–1.68)   
Low and > 31 day 2.42 (2.05–2.84)   
Colon (C18)  ≤ 93  ≤ 126  ≤ 166 
High and ≤ 31 day 
High and > 31 day 1.04 (0.89–1.22) 1.04 (0.88–1.24) 1.08 (0.87–1.33) 
Low and ≤ 31 day 1.27 (1.18–1.36) 1.32 (1.22–1.41) 1.37 (1.26–1.48) 
Low and > 31 day 1.48 (1.30–1.67) 1.53 (1.35–1.73) 1.55 (1.36–1.76) 
Rectal (C19, C20)  ≤ 35   
High and ≤ 31 day   
High and > 31 day 1.20 (1.08–1.34)   
Low and 31 day 1.21 (1.11–1.32)   
Low and > 31 day 1.84 (1.57–2.16)   
Pancreatic (C25)  ≤ 11   
High and 31 day   
High and > 31 day 1.07 (0.84–1.36)   
Low and 31 day 1.21 (1.08–1.36)   
Low and > 31 day 1.60 (1.33–1.92)   
Lung (C33, C34)  ≤ 25  ≤ 67  ≤ 84 
High and 31 day 
High and > 31 day 1.00 (0.89–1.13) 1.04 (0.90–1.20) 1.07 (0.90–1.26) 
Low and 31 day 1.30 (1.20–1.41) 1.43 (1.31–1.55) 1.45 (1.33–1.58) 
Low and > 31 day 1.94 (1.69–2.23) 1.83 (1.62–2.08) 1.82 (1.61–2.07) 
Breast (C50)  ≤ 100  ≤ 201  
High and 31 day  
High and > 31 day 1.21 (0.97–1.51) 1.03 (0.77–1.36)  
Low and 31 day 1.28 (1.13–1.44) 1.17 (1.03–1.33)  
Low and > 31 day 2.66 (2.18–3.24) 2.37 (1.96–2.87)  
Variable aHRc (95% CI) aHRc (95% CI) aHRc (95% CI) 
Type of cancer    
Stomach (C16)  ≤ 17   
High and 31 day   
High and > 31 day 1.01 (0.96–1.05)   
Low and ≤ 31 day 1.54 (1.41–1.68)   
Low and > 31 day 2.42 (2.05–2.84)   
Colon (C18)  ≤ 93  ≤ 126  ≤ 166 
High and ≤ 31 day 
High and > 31 day 1.04 (0.89–1.22) 1.04 (0.88–1.24) 1.08 (0.87–1.33) 
Low and ≤ 31 day 1.27 (1.18–1.36) 1.32 (1.22–1.41) 1.37 (1.26–1.48) 
Low and > 31 day 1.48 (1.30–1.67) 1.53 (1.35–1.73) 1.55 (1.36–1.76) 
Rectal (C19, C20)  ≤ 35   
High and ≤ 31 day   
High and > 31 day 1.20 (1.08–1.34)   
Low and 31 day 1.21 (1.11–1.32)   
Low and > 31 day 1.84 (1.57–2.16)   
Pancreatic (C25)  ≤ 11   
High and 31 day   
High and > 31 day 1.07 (0.84–1.36)   
Low and 31 day 1.21 (1.08–1.36)   
Low and > 31 day 1.60 (1.33–1.92)   
Lung (C33, C34)  ≤ 25  ≤ 67  ≤ 84 
High and 31 day 
High and > 31 day 1.00 (0.89–1.13) 1.04 (0.90–1.20) 1.07 (0.90–1.26) 
Low and 31 day 1.30 (1.20–1.41) 1.43 (1.31–1.55) 1.45 (1.33–1.58) 
Low and > 31 day 1.94 (1.69–2.23) 1.83 (1.62–2.08) 1.82 (1.61–2.07) 
Breast (C50)  ≤ 100  ≤ 201  
High and 31 day  
High and > 31 day 1.21 (0.97–1.51) 1.03 (0.77–1.36)  
Low and 31 day 1.28 (1.13–1.44) 1.17 (1.03–1.33)  
Low and > 31 day 2.66 (2.18–3.24) 2.37 (1.96–2.87)  

aDelay of cancer surgery > 31 day versus 31 day.

bThe bold values mean the criteria for hospital volume category. In sensitivity analyses, we recategorized hospital volume as a binary variable according to the criteria established by the Leapfrog Group for pancreatic resection (≥11 per year [21]) and the potential candidates of volume-based referral for stomach resection (≥17 per year), lung (25 [6], 67 [19], and 84 [5] per year), and breast (100 [3] and 201 [2] > per year).

cAdjusted for age, sex, Charlson scale, hospital type, insurance, radiotherapy, chemotherapy, type of medical care institution and year of diagnosis.

aHR, adjusted hazard ratio; CI, confidence interval.

discussion

In this study of the survival of patients with any of six types of cancer who were treated surgically in Korean hospitals from 2001 through 2005, we showed that the effects of hospital volume and treatment delay varied.

Our finding that risk-adjusted long-term conditional survival was better for patients who underwent surgery at high-volume hospitals than at low- to medium-volume hospitals regardless of cancer site is consistent with several earlier studies [2–7, 19], but not with other studies of patients with lung [24], rectal [25], or gastric cancer [26], wherein hospital volume appeared to have no effect on overall survival.

The relationships between hospital volume and long-term survival after cancer surgery could not be fully explained by volume-related differences in the patient characteristics identified in this study (age, sex, hospital type, type of insurance, type of cancer, and Charlson index). Numerous quality-of-care factors may contribute to improved long-term outcomes at high-volume hospitals, including completeness of resection [27], number of lymph nodes harvested [3], utilization of multimodality therapy (such as administration of adjuvant systemic therapy) [7], more intensive surveillance with follow-up treatment of recurrent cancer [5, 6], and better treatment of comorbidities [6], all good prognostic factors. In the present study, however, addition of adjuvant chemotherapy and radiotherapy to Cox proportional hazards modeling as independent variables did not alter the main outcomes.

Despite the general recommendation that the waiting time from diagnosis to surgery should not exceed 4 weeks [28] or 31 days, we found that 13.6% of surgery patients in Korea waited > 31 days and that patients at high-volume hospitals waited longer for treatment than those in low- to medium-volume hospitals, We showed that many factors, both personal and institutional, influenced wait time, which is consistent with the findings of others [10, 13]. Others have also reported that patients at high-volume hospitals wait longer for treatment than those at low- to medium-volume hospitals [13]. This can be attributed to hospital crowding due to centralization and to an increased travel burden [1, 11].

Although earlier studies reported no effect of therapeutic delay on 5-year survival for breast [14], colorectal [13], lung [29], or pancreatic cancer [12] patients, we did find an effect. Our results were in agreement, however, with another study that looked at wait times of 6 and 12 weeks for breast cancer patients [30]. Thus, the effect of therapeutic delay on prognosis is still not clear, but it differs with type of cancer, stage of disease, and delay duration [13]. Study inconsistencies can be due to variations in study populations, cancer stage, and time frames. Furthermore, some studies are inconclusive due to methodological limitations, such as a failure to perform multivariate analyses with risk adjustment. Surprisingly, there are too few population-based studies with sample sizes large enough to lead to definitive conclusions [13]. The kind of large prospective randomized trial needed to conclusively answer this question would be ethically unacceptable. This study also showed that the effect of treatment delay on survival for patients with each type of cancer differs according to hospital volume, but if the stage is right for surgery, a 1-month delay might put patients at risk for micro metastases.

Our findings support the recommendations of the Canadian Society of Surgical Oncology, the American College of Chest Physicians (ACCP) [31], and the British Thoracic Society [28]. As part of its National Health Service Cancer Plan published in 2000, the UK government introduced cancer waiting time targets, stipulating that the overall time from referral of a patient with suspected cancer to diagnosis and treatment initiation should not exceed 62 days and that patients with confirmed cancer should be treated within 31 days. Furthermore, noncompliance could lead to financial penalties [12]. As of 2007, excellent progress had been reported toward achievement of the 31-day goal [32].

Together with earlier studies, the present study suggests that policymakers should consider strategies for centralization by volume-based referral without subjecting patients to therapeutic delays or unreasonable travel burdens and indicates a need to monitor wait times closely [1, 4, 11]. To accommodate more patients at high-volume hospitals without increasing wait times, either existing high-volume hospitals must increase their capacity or new high-volume hospitals must be developed [1].

Our study had several limitations. First, nonrandomized studies should be interpreted with caution because of the potential for confounding. Second, despite our adjustment of variables known to affect patient outcomes, we could not differentiate the confounding effect of variables such as cancer stage that may have impacted patient outcomes, at least for some of the less frequently carried out operations, such as pancreatic and lung resection. Third, the reason for treatment delay was unknown in these cases. The patients with greater delay may have been more ill or presented with more advanced disease, which required greater preoperative intervention, but these factors could not be controlled in our analysis. Fourth, we cannot exclude the possibility that the threshold for the impact of therapeutic delay on survival might differ with type of cancer and treatment. Evidence for maximum acceptable wait times is not adequate, and further studies are needed. Finally, our definitions of high-volume and medium- to low-volume hospitals were arbitrary and did not necessarily reflect thresholds for optimal performance [6, 7]. Sensitivity analysis based on minimal volume standards, however, supports our conclusions.

Despite its limitations, our study suggests that the combined effect of hospital volume and therapeutic delay on overall survival should be considered when formulating or revising national health policy on quality cancer care. Continuous monitoring of the impact of centralization and treatment delay is crucial for evaluating policy at the national level.

funding

National Cancer Center (1010081).

disclosure

The authors have declared no conflicts of interest.

acknowledgements

We thank Miriam Bloom (SciWrite Biomedical Writing & Editing Services) for professional editing.

references

1
Stitzenberg
KB
Sigurdson
ER
Egleston
BL
, et al.  . 
Centralization of cancer surgery: implications for patient access to optimal care
J Clin Oncol
 , 
2009
, vol. 
27
 (pg. 
4671
-
4678
)
2
Chen
CS
Liu
TC
Lin
HC
Lien
YC
Does high surgeon and hospital surgical volume raise the five-year survival rate for breast cancer? A population-based study
Breast Cancer Res Treat
 , 
2008
, vol. 
110
 (pg. 
349
-
356
)
3
Peltoniemi
P
Peltola
M
Hakulinen
T
, et al.  . 
The effect of hospital volume on the outcome of breast cancer surgery
Ann Surg Oncol
 , 
2011
, vol. 
18
 (pg. 
1684
-
1690
)
4
Schrag
D
Cramer
LD
Bach
PB
, et al.  . 
Influence of hospital procedure volume on outcomes following surgery for colon cancer
JAMA
 , 
2000
, vol. 
284
 (pg. 
3028
-
3035
)
5
Bilimoria
KY
Bentrem
DJ
Feinglass
JM
, et al.  . 
Directing surgical quality improvement initiatives: comparison of perioperative mortality and long-term survival for cancer surgery
J Clin Oncol
 , 
2008
, vol. 
26
 (pg. 
4626
-
4633
)
6
Birkmeyer
JD
Sun
Y
Wong
SL
Stukel
TA
Hospital volume and late survival after cancer surgery
Ann Surg
 , 
2007
, vol. 
245
 (pg. 
777
-
783
)
7
Gooiker
GA
van Gijn
W
Wouters
MW
, et al.  . 
Systematic review and meta-analysis of the volume-outcome relationship in pancreatic surgery
Br J Surg
 , 
2011
, vol. 
98
 (pg. 
485
-
494
)
8
Dudley
RA
Johansen
KL
Brand
R
, et al.  . 
Selective referral to high-volume hospitals: estimating potentially avoidable deaths
JAMA
 , 
2000
, vol. 
283
 (pg. 
1159
-
1166
)
9
Finks
JF
Osborne
NH
Birkmeyer
JD
Trends in hospital volume and operative mortality for high-risk surgery
N Engl J Med
 , 
2011
, vol. 
364
 (pg. 
2128
-
2137
)
10
Simunovic
M
Theriault
ME
Paszat
L
, et al.  . 
Using administrative databases to measure waiting times for patients undergoing major cancer surgery in Ontario, 1993-2000
Can J Surg
 , 
2005
, vol. 
48
 (pg. 
137
-
142
)
11
Birkmeyer
JD
Siewers
AE
Marth
NJ
Goodman
DC
Regionalization of high-risk surgery and implications for patient travel times
JAMA
 , 
2003
, vol. 
290
 (pg. 
2703
-
2708
)
12
Raptis
DA
Fessas
C
Belasyse-Smith
P
Kurzawinski
TR
Clinical presentation and waiting time targets do not affect prognosis in patients with pancreatic cancer
Surgeon
 , 
2010
, vol. 
8
 (pg. 
239
-
246
)
13
Simunovic
M
Rempel
E
Theriault
ME
, et al.  . 
Influence of delays to nonemergent colon cancer surgery on operative mortality, disease-specific survival and overall survival
Can J Surg
 , 
2009
, vol. 
52
 (pg. 
E79
-
E86
)
14
Brazda
A
Estroff
J
Euhus
D
, et al.  . 
Delays in time to treatment and survival impact in breast cancer
Ann Surg Oncol
 , 
2010
, vol. 
17
 
suppl 3
(pg. 
291
-
296
)
15
Yoo
KY
Cancer control activities in the Republic of Korea
Jpn J Clin Oncol
 , 
2008
, vol. 
38
 (pg. 
327
-
333
)
16
Shin
HR
Won
YJ
Jung
KW
, et al.  . 
Nationwide cancer incidence in Korea, 1999∼2001; first result using the national cancer incidence database
Cancer Res Treat
 , 
2005
, vol. 
37
 (pg. 
325
-
331
)
17
International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM)
 , 
2010
Hyattsville, MD
National Center for Health Statistics
18
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
, vol. 
43
 (pg. 
1130
-
1139
)
19
Bach
PB
Cramer
LD
Schrag
D
, et al.  . 
The influence of hospital volume on survival after resection for lung cancer
N Engl J Med
 , 
2001
, vol. 
345
 (pg. 
181
-
188
)
20
Birkmeyer
JD
Siewers
AE
Finlayson
EV
, et al.  . 
Hospital volume and surgical mortality in the United States
N Engl J Med
 , 
2002
, vol. 
346
 (pg. 
1128
-
1137
)
21
Birkmeyer
JD
RAD
 , 
2004
Evidence-Based Hospital Referral Washington, DC
Factsheet
22
Crawford
SC
Davis
JA
Siddiqui
NA
, et al.  . 
The waiting time paradox: population based retrospective study of treatment delay and survival of women with endometrial cancer in Scotland
BMJ
 , 
2002
, vol. 
325
 pg. 
196
 
23
Sharpe
D
Williams
RN
Ubhi
SS
, et al.  . 
The “two-week wait” referral pathway allows prompt treatment but does not improve outcome for patients with oesophago-gastric cancer
Eur J Surg Oncol
 , 
2010
, vol. 
36
 (pg. 
977
-
981
)
24
Osada
H
Yamakoshi
E
Hospital volume and surgical outcomes of lung cancer in Japan
Gen Thorac Cardiovasc Surg
 , 
2007
, vol. 
55
 (pg. 
360
-
365
)
25
Kressner
M
Bohe
M
Cedermark
B
, et al.  . 
The impact of hospital volume on surgical outcome in patients with rectal cancer
Dis Colon Rectum
 , 
2009
, vol. 
52
 (pg. 
1542
-
1549
)
26
Xirasagar
S
Lien
YC
Lin
HC
, et al.  . 
Procedure volume of gastric cancer resections versus 5-year survival
Eur J Surg Oncol
 , 
2008
, vol. 
34
 (pg. 
23
-
29
)
27
Bilimoria
KY
Talamonti
MS
Sener
SF
, et al.  . 
Effect of hospital volume on margin status after pancreaticoduodenectomy for cancer
J Am Coll Surg
 , 
2008
, vol. 
207
 (pg. 
510
-
519
)
28
 
British Thoracic Society. Recommendations to respiratory physicians for organising the care of patients with lung cancer. The Lung Cancer Working Party of the British Thoracic Society Standards of Care Committee. Thorax 1998; 53 (suppl 1): S1–S8
29
Myrdal
G
Lambe
M
Hillerdal
G
, et al.  . 
Effect of delays on prognosis in patients with non-small cell lung cancer
Thorax
 , 
2004
, vol. 
59
 (pg. 
45
-
49
)
30
Hershman
DL
Wang
X
McBride
R
, et al.  . 
Delay in initiating adjuvant radiotherapy following breast conservation surgery and its impact on survival
Int J Radiat Oncol Biol Phys
 , 
2006
, vol. 
65
 (pg. 
1353
-
1360
)
31
Alberts
WM
Bepler
G
Hazelton
T
, et al.  . 
Lung cancer. Practice organization
Chest
 , 
2003
, vol. 
123
 (pg. 
332S
-
337S
)
32
 
National Health Service. Cancer Reform Strategy. UK: Department of Health 2007

Author notes

DSH contributed equally to this work as corresponding author.