© 2007 Foundation for Promotion of Cancer Research
Hospital Procedure Volume and Survival of Cancer Patients in Osaka, Japan: A Population-based Study with Latest Cases
1 Department of Cancer Control and Statistics, Osaka Medical Center for Cancer and Cardiovascular Diseases, Osaka
2 Statistics and Cancer Control Division, National Cancer Center, Tokyo, Japan
For reprints and all correspondence: Akiko Ioka, Department of Cancer Control and Statistics, Osaka Medical Center for Cancer and Cardiovascular Diseases, 1-3-3 Nakamichi, Higashinari-ku, Osaka 537-8511, Japan. E-mail: akiko3{at}gol.com
Received October 2, 2006; accepted February 24, 2007
 |
Abstract |
|---|
 TOP Abstract INTRODUCTIONMETHODSResultsDiscussionConflict of interest statementReferences |
|---|
Background: Recent studies reported that hospital procedure volume (i.e. volume of patients per hospital receiving a particular treatment)was directly proportional to cancer survival; however the degree of association might be different according to the primary tumor site, extent of disease and year of diagnosis. We performed a systematical examination of survi vals by hospital procedure volume according to the primary site with inclusion of latest cases in Osaka, Japan.
Methods: Individual data on reported cancer cases with active follow-up information and diagnosis between 1994 and 1998 were retrieved from Osaka Cancer Registry's database. The analysed primary sites included oesophagus, stomach, large bowel, liver, gall bladder, pancreas, lung, breast, uterus, ovary, prostate, bladder and lymphoma. Hospitals were ranked as high-, medium-, low- and very low-volume hospitals for every primary site by dividing the number of cancer patients who received treatment in hospitals into four quartiles.
Results: The primary sites could be classified into three categories based on the association between hospital procedure volume and cancer survival: In type 1, a better survival was associated with a higher procedure volume as for oesophagus, liver, lung, ovary, prostate, or lymphoma; in type 2, a better survival was associated with a higher procedure volume but there was no significant difference in survival between high- and medium-volume hospitals as for uterus; and in type 3, there was no significant difference in survival among high-, medium- and low-volume hospitals as for stomach, large bowel, gall bladder, pancreas, breast, or bladder sites.
Conclusions: A higher procedure volume was generally associated with a better survival; however, this association could be classified into three types according to the primary site.
Key Words: hospital procedure volume • survival • cancer • primary site
|
INTRODUCTION |
|---|
|
TOPAbstractINTRODUCTIONMETHODSResultsDiscussionConflict of interest statementReferences |
|---|
Recent studies have suggested that hospital procedure volume is directly proportional to cancer survival (1–4) however, the degree of association might be different according to the primary site, extent of disease and year of diagnosis as suggested partly in our studies (5–8). We have thus tried to clarify the association between hospital procedure volume (i.e. the number of patients who received treatment) and survival systematically according to the primary site using the latest data of the Osaka Cancer Registry (OCR).
|
METHODS |
|---|
|
TOPAbstractINTRODUCTIONMETHODSResultsDiscussionConflict of interest statementReferences |
|---|
Data Sources
Individual data on reported cancer cases with active follow-up information and a diagnosis during 1994–98 were retrieved from the database of the OCR, which has been operating since December 1962 and covers Osaka Prefecture with a population of 8.8 million (2000 census) (9,10). Criteria for the analysis were as follows:
- In the case of multiple tumors, only the first was included.
- Cases diagnosed as carcinoma in situ or mucosal cancer of colon and rectum were excluded.
- Primary sites were oesophagus (C15, ICD Tenth Revision), stomach (C16), large bowel (C18–C21), liver (C22), gall bladder, etc. (C23–C24), pancreas (C25), trachea, bronchus and lung (C33–C34), breast (C50), uterus (C53–C55), ovary (C56), prostate (C61), bladder (C67) and lymphoma (C81–C90, C96).
In the processing of OCR's data for cancer statistics, primary facilities for treatment of each cancer were determined and coded in the following order: surgery, radiotherapy, transarterial embolization, ethanol injection, chemotherapy, immunotherapy, hormone therapy, laser therapy and thermotherapy. Hospitals were ranked as high-, medium-, low- and very low-volume hospitals for every site by dividing the number of cancer patients who received treatment in hospitals, excluding clinics and unknown, into four quartiles. The cancer stage at diagnosis was classified into the following three categories:
- Localized: cancer is confined to the original organ.
- Regional: cancer has spread to regional lymph nodes and/or to immediately adjacent tissues.
- Distant: cancer has metastases to distant organs.
Statistical Analysis
A cumulative survival was estimated using the Kaplan–Meier method for each category of hospital procedure volume. Survival time was computed from the date of first diagnosis to the end-point, defined as death from any cause. Closing date was defined as the date 5 years after the first diagnosis. The relative 5-year survival (11,12) was calculated as the ratio of the observed survival to the expected survival estimated using the survival probability of similar subjects in the general population of Japan with respect to sex, age and calendar year at diagnosis. The prognostic factors were adjusted using the Cox proportional hazards regression model during the 5 years following diagnosis. In this analysis, independent variables were sex, age, cancer stage (localized, regional, distant and unknown), and hospital procedure volume (high-, medium-, low- and very low-procedure volume). Differences were considered as statistically significant if P values were less than 0.05 by two-sided test. The statistical package software STATA (13) was used for data management and statistical analysis.
|
Results |
|---|
|
TOPAbstractINTRODUCTIONMETHODSResultsDiscussionConflict of interest statementReferences |
|---|
Tables 1
–13 illustrate the relative 5-year survival by hospital procedure volume according to the primary site. In stomach, large bowel and breast cancer, the survival in medium-/low-volume hospitals was almost the same as that in high-volume hospitals, however, in other cancers, the survival increased with increasing hospital procedure volume. The associations between the hospital procedure volume and adjusted hazard ratio (HR) were classified into three types (Fig. 1): type 1 showed a higher procedure volume–lower HR association for oesophagus, liver, lung, ovary, prostate and lymphoma; type 2 showed a higher procedure volume–lower HR association but equivalent HRs between high- and medium-volume hospitals for the uterus; and type 3 showed almost equivalent HRs among high-, medium- and low-volume hospitals for stomach, large bowel, gall bladder, pancreas, breast and bladder.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Discussion |
|---|
|
TOPAbstractINTRODUCTIONMETHODSResultsDiscussionConflict of interest statementReferences |
|---|
A population-based study with recent data from OCR suggested a significant association between hospital procedure volume and survival. Three types of association were observed according to the primary site. This study is unique and valuable for the use of population-based data with active follow-ups, although several limitations were inherent.
These relationships might have been confounded by treatment modalities: operable cases might have been treated by higher volume hospitals even in the same stage and primary site, which lead to spurious associations. We have already examined relationships between hospital surgical volume and survival for cancers of the stomach, breast, uterus and ovary, which all showed findings similar to the present study. Because treatment procedures closely correlate with cancer stage, we analysed the associations according to the stage and/or with adjustment for the stage. However, the observed associations might be still confounded by treatment modalities.
Hospitals were classified into four categories by dividing the number of cancer patients who received treatments in hospitals into four quartiles, as high-, medium-, low- and very low-volume hospitals based on a bigger number of patients per hospital-month. In this study however, the number of patients per hospital-month was very few in several primary sites even in high-volume hospitals. Therefore, the interpretation of the high-volume category should be made carefully because it may not necessarily mean physicians and medical teams belonging to high-volume hospitals had lots of experience of cancer treatments.
We suggested that the primary sites were classified into three categories based on the association between hospital procedure volume and cancer survival, which was supported by partition clustering method of medians: when a high-volume hospital was taken as a reference, adjusted HRs were categorized into four clusters (i.e. almost the same HRs as 1.0 with median 1.0, much higher HRs than 1.0 with median 2.1, and lower/higher intermediate HRs between these with median 1.3/1.6) using this method. The primary sites were, then, reasonably classifiable into types 1–3 when we considered combinations of these clusters of HRs and 95% CI of HRs.
In lung, liver and prostate cancers belonging to type1, a much higher survival in high-volume hospitals might have been influenced by stage migration as well as insufficient adjustment for cancer stage distribution: for example, in lung cancer at the localized stage, the survival in high-volume hospitals was double that in very low-volume hospitals. In gall bladder and pancreas cancers belonging to type3, the 5-year survival might have been too long to evaluate associations between hospital procedure volume and survival, as well as in the other cancers with distant metastases.
Some other limitations should be kept in mind in this study. Among the patients' characteristics, we only took sex and age into consideration. We should have also considered the prevalence of co-morbidities and difference of socioeconomic factors, and so on. In addition, as we mentioned in prior studies, we should have taken into consideration the completeness of reporting to the cancer registry and quality of information on treatment.
Despite the limitations mentioned above, our analysis is now one of the few approaches to clarify the association between hospital procedure volume and survival. The study results suggest that there were three types of relationships between hospital procedure volume and cancer survival. The authors consider that epidemiology data like this would be very important for the planning and execution of effective cancer control programs in Osaka where there are many hospitals.
|
Conflict of interest statement |
|---|
|
TOPAbstractINTRODUCTIONMETHODSResultsDiscussionConflict of interest statementReferences |
|---|
None declared.
|
References |
|---|
|
TOPAbstractINTRODUCTIONMETHODSResultsDiscussionConflict of interest statementReferences |
|---|
1 Killeen SD, O'Sullivan MJ, Coffey JC, Kirwan WO, Redmond HP. Provider volume and outcomes for oncological procedures. Br J Surg (2005) 92:389–402.[CrossRef][Web of Science][Medline]
2 Hillner BE. Is cancer care best at high-volume providers? Curr Oncol Rep (2001) 3:404–9.[Medline]
3 Birkmeyer JD, Siewers AE, Finlayson EV, Stukel TA, Lucas FL, Batista I, et al. Hospital volume and surgical mortality in the United States. N Engl J Med (2002) 346:1128–37.
4 Begg CB, Cramer LD, Hoskins WJ, Brennan MF. Impact of hospital volume on operative mortality for major cancer surgery. JAMA (1998) 280:1747–51.
5 Ioka A, Tsukuma H, Ajiki W, Oshima A. Influence of hospital procedure volume on uterine cancer survival in Osaka, Japan. Cancer Sci (2005) 96:689–94.[CrossRef][Medline]
6 Ioka A, Tsukuma H, Ajiki W, Oshima A. Influence of hospital procedure volume on ovarian cancer survival in Japan, a country with low incidence of ovarian cancer. Cancer Sci (2004) 95:233–7.[CrossRef][Medline]
7 Nomura E, Tsukuma H, Ajiki W, Oshima A. Population-based study of relationship between hospital surgical volume and 5-year survival of stomach cancer patients in Osaka, Japan. Cancer Sci (2003) 94:998–1002.[CrossRef][Medline]
8 Nomura E, Tsukuma H, Ajiki W, Ishikawa O, Oshima A. Population-based study of the relationship between hospital surgical volume and 10-year survival of breast cancer patients in Osaka, Japan. Cancer Sci (2006) 97:618–22.[CrossRef][Medline]
9 Ajiki W, Tsukuma H, Oshima A. Trends in cancer incidence and survival in Osaka. In: Cancer Mortality and Morbidity Statistics: Japan and the World—2004—Tajima K, Oshima A, Kuroishi T, eds. (2004) Tokyo: Japan: Sci Soc Press. 137–63. Gann Monograph on Cancer Research No. 51.
10 Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas DB. Cancer Incidence in Five Continents. (2002) Vol. VIII. Lyon, France: International Agency for Research on Cancer. IARC Scientific Publ. No. 155.
11 Ajiki W, Matsuda T, Sato Y, Fujita M, Yamazaki S, Murakami R, et al. Standard method of calculating relative survival rates in population-based cancer registries—an investigation using stomach cancer patients. Jpn J Cancer Cli (1997) 43:1005–14.
12 Estev J, Benhamou E, Raymond L. Statistical Methods in Cancer Research. (1994) Vol. IV. Lyon, France: Descriptive Epidemiology. 231–45. IARC Scientific Publ. No. 128, International Agency for Research on Cancer.
13 StataCorp. Stata Statistical Software: Release 8.0. (2003) Texas: US Stata Corporation.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Tsutsui, Y. Ohno, J. Hara, Y. Ito, and H. Tsukuma Trends of Centralization of Childhood Cancer Treatment Between 1975 and 2002 in Osaka, Japan Jpn. J. Clin. Oncol., February 1, 2009; 39(2): 127 - 131. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||