SEARCH

SEARCH BY CITATION

Keywords:

  • ovarian cancer;
  • disparities;
  • regional variation;
  • epidemiology;
  • chemotherapy

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

BACKGROUND:

Regional differences in health services can point to disparities in access to healthcare. The authors performed a population-based cohort study to examine differences in ovarian cancer treatment and mortality according to geographic region.

METHODS:

The Surveillance, Epidemiology, and End Results (SEER)-Medicare database was used to identify 4589 women aged ≥65 years with ovarian cancer diagnosed between 1998 and 2002 who had Medicare claims filed from 1998 to 2005. Hospital Referral Region (HRR) was assigned according to patient zip code. The authors calculated the proportion of women in each HRR who underwent cancer-directed surgery. With HRR as the predictor of interest, mortality and the receipt of cancer-directed surgery were described in multivariate analyses.

RESULTS:

Among 4589 women with ovarian cancer, 3286 underwent cancer-directed surgery. The receipt of cancer-directed surgery varied by HRR (range, 53%-88%). Women were less likely to undergo cancer-directed surgery if they were older, nonwhite, had higher stage disease, or had more comorbidities. For example, white women were more likely to undergo such surgery (odds ratio, 1.41; 95% confidence interval, 1.10-1.82) compared with all nonwhite women. HRR was a significant predictor of cancer-directed surgery (P = .01). A significant correlation was observed between HRR and all-cause mortality (P = .02); however, after adjusting for cancer-directed surgery, that correlation was no longer significant (P = .10).

CONCLUSIONS:

There was regional variation in mortality among Medicare recipients with ovarian cancer, and access to cancer-directed surgery explained some of that variation. Improving access to high-quality cancer surgery for ovarian cancer may improve outcomes, particularly for minorities and for older women. Cancer 2010. © 2010 American Cancer Society.

Ovarian cancer is the fourth most common cause of death from malignancies among US women, with 21,650 new cases and 15,520 deaths expected in 2008.1 The majority of women (67%-69%) are diagnosed with late-stage disease.2 Although disease stage at diagnosis is the most important predictor of survival in ovarian cancer, even women with advanced disease can benefit from surgery.2 Initial management of ovarian cancer includes appropriate surgery for accurate staging and optimal cytoreduction (ie, debulking).3 There is evidence of variation in treatment for ovarian cancer, specifically, older women with late-stage disease are less likely to receive recommended surgery and chemotherapy.4, 5 Appropriate surgical treatment6 occurs more commonly if the treating physician is a gynecologic oncologist as opposed to a general gynecologist or a general surgeon.7-11 In an analysis of Surveillance, Epidemiology, and End Results (SEER)-Medicare data by Earle et al, treatment by a gynecologic oncologist improved overall survival, and the availability of gynecologic oncologists was the strongest predictor of receiving care by this type of specialist.11

Geographic variation in the delivery of health services is an important factor in understanding the use of, quality, disparities in, and access to care for a variety of healthcare services.12-14 Earlier studies reported geographic variation in cancer screening, surveillance, and treatment for selected cancers. To our knowledge, there no studies to date have assessed geographic variation in the treatment of ovarian cancer. In this analysis, we describe regional variation in the receipt of cancer-directed surgery and chemotherapy for ovarian cancer in the Medicare population as well as regional variations in all-cause mortality and disease-specific mortality.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

Data Sources

Data for this analysis came from the linkage of individuals in the Surveillance, Epidemiology, and End Results (SEER) registries with their healthcare claims reported to the Medicare program. The population-based registries that have participated in the SEER Program represent defined geographic areas and have changed over time. The registries that we included in the current analysis range from 14% of the US population (1998-2000) to 26% of the US population (2001-2005).15 An estimated 97% of incident cancer cases are captured in the SEER regions,16 which are representative of the US population.17 For each reported malignancy, the SEER registries collect data on age at diagnosis, sex, race and ethnicity, stage, histologic type, month and year of diagnosis, whether cancer-directed surgery was completed, and the date and cause of death. SEER registrars receive training and follow detailed manuals for tumor reporting (available at: http://seer.cancer.gov/registrars/ accessed October 13, 2009). Sociodemographic data from the 2000 census are linked with the cases at the census level.

The Medicare data include claims for inpatient and outpatient care, physician services, and laboratory testing, among others. SEER and Medicare data are linked periodically for research purposes to ascertain treatment and outcomes not captured by SEER, and the data have a match rate of 94%.18 The linked dataset for the current analysis included claims from January 1, 1998 through December 31, 2005.

Approval was obtained from the Human Subjects Committee at Maine Medical Center for the current data analysis. Because the SEER-Medicare data are deidentified and are based on registry data, it is not possible to seek informed consent from each participant; therefore, we requested a waiver of consent. Because we required the use of each patient's zip code for this study, we obtained permission from each of the SEER cancer registrars involved before we were given access to the data. No other patient identifiers were included with the data.

Cohort Definition

The study patients included women in the SEER-Medicare data who were diagnosed with epithelial ovarian cancer at age ≥65 years between 1998 and 2002 (inclusive) while living in a SEER area. To identify comorbidities that were present before diagnosis, women in the cohort had to have been enrolled in Medicare at least 6 months before diagnosis. We excluded women who had a history of prior malignancy, had borderline ovarian cancer pathology, had disease that was not pathologically confirmed, or were diagnosed on death certificate or autopsy only. We examined frequencies of borderline pathology type and did not observe any differences according to geographic region. All histologic subtypes of epithelial tumors (after excluding borderline tumors) were considered together in this analysis of registry data. Women were included in the analysis if they were eligible for both Medicare Part A and Part B and if they were not enrolled in a health maintenance organization at diagnosis for the entire follow-up period. Patients were followed to ascertain outcomes from their claims using data through 2005 (Fig. 1).

thumbnail image

Figure 1. This chart describes the patients with ovarian cancer who were included in the study cohort (1998-2002). SEER indicates the National Cancer Institute's Surveillance, Epidemiology, and End Results Program.

Download figure to PowerPoint

Definitions of Explanatory Variables

Geographic areas

To assess geographic variation, we used the Hospital Referral Regions (HRRs) and Hospital Service Areas (HSAs) as defined by Wennberg and Cooper in the Dartmouth Atlas of Health Care.19 HRRs represent markets for tertiary medical care, whereas HSAs represent local markets for community-based inpatient care. HSAs are assigned to an HRR based on where the majority of patients within an HSA went for major cardiovascular procedures. On the basis of these original definitions, there were 3436 HSAs and 306 HRRs in the United States. The details of definitions and development of these regions (based on acute care hospitalizations of the Medicare population in 1992 and 1993) are described in detail in the Dartmouth Atlas of Health Care. They have been used by health services researchers for over a decade to describe regional variation in the use of a variety of care processes.

Women who were included in our analysis were assigned to 1 of 74 HRRs based on the patient's zip code of residence at the time of diagnosis. We do not have all 306 HRRs, because the SEER Program is limited to specific states and because we suppressed data for HRRs if there were not at least 10 ovarian cancer cases. Each HRR was converted to a dummy variable. Salt Lake City, Utah was used as the referent HRR, because we observed that women with epithelial ovarian cancer in this area had the median survival for the women in our cohort after diagnosis.

Patient characteristics

Sociodemographic variables that were included in the current analysis were patient age and race, which we defined as white or nonwhite based on SEER data. Disease stage using the American Joint Committee on Cancer (AJCC) classification was obtained from the SEER record for each patient. Comorbidities were ascertained using inpatient and physician visit claims for diagnostic billing codes20 during the 6 months before the cancer diagnosis using the implementation described by Deyo et al21 of the Charlson comorbidity score.22 Each patient's state/county of residence was mapped, and the map was used to identify whether the woman resided in an urban or rural location.23 In addition to individual-level variables, we used census data for the median household income within the zip code of the patient's residence as a proxy to measure the patient's economic status. Population density of obstetrician/gynecologists by HRR was obtained from the 1996 Dartmouth Atlas of Health Care and was explored as an additional explanatory variable in some analyses.

Surgical care

We defined cancer-directed surgery as resection of the primary tumor (with or without hysterectomy and/or debulking). Surgeries were identified by reviewing Medicare claims for inpatient services (Medicare Provider Analysis and Review [MEDPAR]), physician bills, and outpatient services. We searched claims for 90 days before the SEER diagnosis date and 365 days after that date for evidence of surgery. Relevant International Classification of Diseases, ninth revision, clinical modification (ICD-9-CM) codes and American Medical Association Current Procedural Terminology (CPT) codes for initial and subsequent ovarian cancer surgeries, treatments, and outcomes were assessed as they developed by an expert panel of gynecologists and gynecologic oncologists, as described in the text and appendices of an article by Earle et al.11 Analyses of 30-day surgical mortality were based on the date of cancer-directed surgery from the claims. An additional variable from SEER indicating why cancer-directed surgery was not performed (including categories for not recommending surgery and patient refusal) was used to consider patients who never underwent cancer-directed surgery. We used a variable from the Physician/Supplier file that indicated the specialty of the treating physician to identify whether patients who did not undergo surgery ever were seen by a gynecologic oncologist, gynecologist, or general surgeon.

Chemotherapy

By using the same claims files described above to ascertain surgery, we reviewed claims for evidence of chemotherapy administration in the 6 months after diagnosis. We used the ICD-9 procedure code 99.25; ICD-0 diagnosis codes V58.1, V66.2, and V67.2 in MEDPAR; CPT codes 96, 400 through 496, and 549; and Hospital Common Procedure Coding System codes J9000 through J9999 and Q0083 through Q0085 in the Physician/Supplier and Outpatient files.

Outcomes

Receipt of cancer-directed surgery and chemotherapy were the outcomes of interest in our analyses of variation in treatment according to geographic region. In analyses of mortality, we were interested primarily in overall and cancer-specific mortality. Overall survival was calculated from the 15th of the month of diagnosis to the date of death from any cause (SEER collects only month and year of diagnosis). Patients were censored if they were alive at their last available Medicare coverage month. For analyses of disease-specific mortality, SEER data were used to determine cause of death and to assign patients to death from ovarian cancer or from other conditions. We also assessed 30-day surgical mortality for the women who underwent cancer-directed surgery.

Statistical Analysis

We used logistic regression to perform analyses for receipt of cancer-directed surgery and chemotherapy. We adjusted first for clinical factors like age, stage at diagnosis, and comorbidity score (Model 1). Then, we added demographic characteristics to the model, including race, income, and education (according to the census tract for patient residence) (Model 2). Finally, we added indicators for each HRR to the model (Model 3). To obtain a P value for HRR effect, we compared the −2 log-likelihood for a full model without HRR included to a model with HRR added and used a chi-square test with 73 degrees of freedom (the number of HRR indicator variables minus 1). In additional models (without HRR in the model), we separately explored SEER registry area and a variable indicating urban versus rural area of residence. In addition, we evaluated the population density of obstetrician/gynecologists by HRR in the model and again without HRR in the model.

We used Cox proportional hazards models to evaluate the associations between HRR and all-cause mortality and between HRR and disease-specific mortality. In these models, we adjusted for age, cancer stage, comorbidity, race, income, and education, as in the models described above. Indicator variables for HRR of residence were added to fully adjusted models. Also as in the models described above, the HRR indicating Salt Lake City, Utah was used as the referent HRR.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

Among 4589 women who were eligible for analyses of mortality, 3286 women (71.6%) underwent cancer-directed surgery (Fig. 1), and 2976 women (64.8%) received chemotherapy. Nearly 30% of women (1370 of 4589) were aged ≥80 years, and most women (4083 of 4589; 89%) were white (Table 1). Seventy-five percent of the women had stage III or IV disease at presentation.

Table 1. Characteristics of the 4589 Women With Epithelial Ovarian Cancer in the Surveillance, Epidemiology, and End Results-Medicare Database Who Met Eligibility Criteria, 1998-2002
CharacteristicNo. of Patients (%)
  • FIGO indicates International Federation of Gynecology and Obstetrics.

  • a

    Charlson cormorbidity scores were based on 6 months of claims data.

Age, y
 65-69926 (20.2)
 70-741120 (24.4)
 75-791173 (25.6)
 ≥801370 (29.9)
Race
 White4083 (89)
 Black270 (5.9)
 Other236 (5.1)
FIGO stage
 I598 (13)
 II338 (7.4)
 III1964 (42.8)
 IV1474 (32.1)
 Missing215 (4.7)
Charlson comorbidity scorea
 03462 (75.4)
 1769 (168)
 ≥2358 (7.8)
Percentage of individuals aged ≥25 y with <12 y education in census tract 
 ≥20%1524 (33.2)
 10-19.9%1607 (35)
 <10%1458 (31.8)
Median income for census tract
 <$40,0001621 (35.3)
 $40,00-$58,0001571 (34.2)
 >$58,0001397 (30.4)
Underwent cancer-directed surgery3286 (71.6)
Received chemotherapy2976 (64.8)

Figure 2 describes variation in cancer-directed surgery for ovarian cancer across the HRRs assigned to SEER areas. Such surgery varied from 53% to 88% (see Fig. 2) across HRRs. Among the 1303 women who did not undergo cancer-directed surgery, 866 women had a least 1 visit with a surgeon (general surgery, gynecology, or gynecologic oncology). For 758 of the 1303 women (58.2%) who did not undergo such surgery, it was not recommended (as reported to the SEER registrars). An additional 62 women had surgery recommended to them but declined.

thumbnail image

Figure 2. This map illustrates the percentage of patients with ovarian cancer who underwent cancer-directed surgery (1998-2002). SEER indicates the National Cancer Institute's Surveillance, Epidemiology, and End Results Program.

Download figure to PowerPoint

Multivariate analyses of cancer-directed surgery (Table 2) indicated that the most elderly women (aged ≥80 years) were least likely to undergo such surgery (odds ratio [OR], 0.16; 95% confidence interval [CI], 0.13-0.20) compared with women ages 65 to 69 years. Women with stage IV cancers were much less likely (OR, 0.12; 95% CI, 0.09-0.16), and whites were more likely (OR, 1.41; 95% CI, 1.10-1.82) to undergo such surgery. Demographic variables (income and education by census tract) were not strongly associated with undergoing cancer-directed surgery, nor was the percentage of non-English speakers according to census tract data (results not shown). In fully adjusted models, HRR was a significant predictor of undergoing cancer-directed surgery (P = .01). Models examining SEER registry and rurality of area of residence as predictors (without HRR in the model) did not reveal strong or consistent associations. In addition, we did not observe an association between population density of obstetrician/gynecologists and receipt of cancer-directed surgery.

Table 2. Factors Associated With the Odds of Undergoing Cancer-Directed Surgery Among 4589 Women With Epithelial Ovarian Cancer (3286 who Underwent Surgery) in the Surveillance, Epidemiology, and End Results-Medicare Database, 1998-2002
VariableModel 1aModel 2bModel 3c
OR (95% CI)POR (95% CI)POR (95% CI)P
  • OR indicts odds ratio; CI, confidence interval, Ref, referent category; HRR, Hospital Referral Region.

  • a

    Model 1 included age, stage, and Charlson comorbidity score.

  • b

    Model 2 included age, stage, Charlson comorbidity score, race, income, and education.

  • c

    Model 3 included age, stage, Charlson comorbidity score, race, income, education, and indicator variables for each individual HRR.

  • d

    The lowest tertile of median income for census tract, annual income <$40,000; middle tertile, $40,000-$57,999; highest tertile, ≥$58,000.

Age, y
 65-691.0 (Ref) 1.0 (Ref) 1.0 (Ref) 
 70-740.70 (0.54-0.90).0050.70 (0.54-0.90).0050.70 (0.54-0.91).008
 75-790.48 (0.38-0.62)<.00010.47 (0.37-0.60)<.00010.47 (0.37-0.60)<.0001
 ≥800.17 (0.14-0.22)<.00010.17 (0.13-0.21)<.00010.16 (0.13-0.20)<.0001
Stage at diagnosis
 I1.0 (Ref) 1.0 (Ref) 1.0 (Ref) 
 II0.59 (0.40-0.88)0.0080.60 (0.40-0.88)0.0090.57 (0.38-0.86).002
 III0.53 (0.40-0.71)<.00010.53 (0.40-0.70)<.00010.53 (0.39-0.71)<.0001
 IV0.13 (0.10-0.17)<.00010.13 (0.10-0.17)<.00010.12 (0.09-0.16)<.0001
Comorbidity score
 01.0 (Ref) 1.0 (Ref) 1.0 (Ref) 
 10.59 (0.48-0.70)<.00010.59 (0.49-0.71)<.00010.56 (0.46-0.68)<.0001
 ≥20.38 (0.30-0.49)<.00010.40 (0.31-0.51)<.00010.39 (0.30-0.50)<.0001
White race vs nonwhite  1.46 (1.16-1.83).0011.41 (1.10-1.82).007
Median income for census tractc
 Lowest tertile  1.0 (Ref) 1.0 (Ref) 
 Middle tertile  1.11 (0.92-1.33).291.18 (0.96-1.46).12
 Top tertile  1.01 (0.80-1.27).931.04 (0.79-1.38).77
Percentage of adults aged 25 y with <12 y of education in census tractd
 Lowest tertile  1.0 (Ref) 1.0 (Ref) 
 Middle tertile  0.97 (0.80-1.17).750.98 (0.80-1.21).87
 Top tertile  1.22 (0.96-1.55).101.30 (1.00-1.70).05
HRR effect in model     .01

Chemotherapy use for ovarian cancer also varied across HRRs within the SEER areas, from 48% to 93% (Fig. 3). Analyses predicting the receipt of chemotherapy (results not shown) revealed similarly that the most elderly women were least likely to receive chemotherapy as were nonwhites, women with more comorbidities, and women with stage IV cancers. HRR was not statistically significant as a predictor of receiving chemotherapy in fully adjusted models (P = .29). It is noteworthy that stage I was used as the reference category for disease stage, and these women would not necessarily be expected to receive chemotherapy; however, 293 of 598 women (49%) with stage I disease did receive chemotherapy.

thumbnail image

Figure 3. This map illustrates the percentage of patients with ovarian cancer who received chemotherapy (1998-2002). SEER indicates the National Cancer Institute's Surveillance, Epidemiology, and End Results Program.

Download figure to PowerPoint

In Cox proportional hazards models of all-cause mortality, we also observed higher mortality among the most elderly women, those with advanced-stage disease and more comorbidities, and nonwhites. HRR in the fully adjusted model was a significant predictor of all-cause mortality (P = .02). However, when we added cancer-directed surgery to the model, HRR no longer was statistically significant (P = .10). In analyses of disease-specific mortality, we observed similar patterns, although they were attenuated given the smaller number of outcomes (specifically, deaths from ovarian cancer). Figure 3 illustrates variation in the percentage of women who died within 2 years according to HRR, which ranged from 27% to 65%.

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

There has been little published information on regional variation in treatment for ovarian cancer. National Comprehensive Cancer Network guidelines recommend that physicians consider cancer-directed surgery for all women with ovarian cancer, regardless of stage.3 We observed that there was significant variation in the use of cancer-directed surgery between HRRs. This variation in surgical rates persisted even after fully adjusting our models for clinical and sociodemographic factors. The areas with the lowest rates of cancer-directed surgery, as illustrated in Figure 2, tended to be in more remote locations. If women are not being offered surgical referral after diagnosis in some regions of the United States, then they cannot realize the benefit of such surgery.

Our current findings related to chemotherapy were in contrast to our findings related to cancer-directed surgery. Although there was considerable observed variation by HRR in the receipt of chemotherapy for ovarian cancer, this variation no longer was important after the models were adjusted for clinical and sociodemographic factors. Our findings with regard to chemotherapy were similar to those from another analysis that used SEER-Medicare data (through 2001) to demonstrate variation in the use of chemotherapy for women with ovarian cancer in the Medicare population.24 In that earlier study, the authors noted that clinical characteristics (stage at diagnosis, age, and comorbidities) were much more important than HRR in explaining chemotherapy use. However, those authors did report a range from 33% to 67% in the proportion of women who received chemotherapy across hospital referral regions in their analysis. We observed a similarly broad range for the receipt of chemotherapy across HRR, although the percentage of women who received chemotherapy, 48% to 93%, was much higher than that reported in the earlier study. These findings likely reflect the increased use of chemotherapy for ovarian cancer over the past several years. Modest observed differences in all-cause mortality and 2-year mortality according to HRR in our study no longer were important after adjusting for the receipt of cancer-directed surgery.

The current analysis has several limitations. The use of claims to determine comorbidities is likely to result in under ascertainment. However, this would not be expected to differ according to geographic region. Because we relied on claims data, we may have missed relevant clinical data that also would explain some of the variation in cancer-directed surgery across HRRs. Some women may have been offered appropriate surgical treatment but declined such therapy in favor of supportive care only, particularly those who may be deemed poor surgical candidates. We did not have access to information on the density of gynecologic oncologists in the HRRs to further refine our analysis of care by physician availability. We cannot comment on women who received their care in health maintenance organizations, who may have had different access to care than women in fee-for-service Medicare populations. We relied on local pathologist reports from the treating facilities for accurate pathology data, which, in turn, are collected by SEER registrars. Because it may be difficult to establish borderline pathology in ovarian cancer, the inclusion of a small number of borderline tumors in the analysis may have occurred inadvertently. However, we did not observe any evidence of variation in the proportion of tumors with borderline pathology by HRR (and borderline tumors were excluded in our analysis), so any misclassification of pathology would not be expected to affect our results. Finally, we cannot comment on the entire Medicare population, because we were limited to women residing within the SEER areas, although the SEER areas represent a large portion of the US population. It is noteworthy that an analysis of a nationally representative sample of women with ovarian cancer in the Medicare population would be limited by a lack of staging information, which is a particular strength of the SEER data.

The current study demonstrates that variation in the receipt of cancer-directed surgery and chemotherapy for ovarian cancer is present among women in the Medicare population living within SEER areas and that differences in surgical rates explain much of the observed variation in outcomes. Because surgery is centrally important in staging and care for these patients, this variation points to substantial room for improvement. Similarly, because chemotherapy also is an important part of therapy for women with stage III and IV ovarian cancer, our findings suggest opportunities for improvement. Even among older women and women with comorbidities, in almost all cases, chemotherapy can be tailored effectively to the individual patient. Women and their providers need to be aware that cancer-directed surgery is important across all stages and that chemotherapy it is an option for patients with stage III and IV cancers. Providers should make appropriate referrals to gynecologic oncologists if this specialized care is not available locally.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

The authors acknowledge key programming input from Andrea Siewers and Kim Murray and assistance with mapping from Laura Kuhl, all at the Center for Outcomes Research and Evaluation, Maine Medical Center. This study used the linked SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the Applied Research Program, NCI; the Office of Research, Development and Information, CMS; Information Management Services (IMS), Inc.; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database

CONFLICT OF INTEREST DISCLOSURES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

Supported by Research funding from Maine Medical Center (to Dr. Fairfield). The collection of the California cancer incidence data used in this study was supported by the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885; the National Cancer Institute's Surveillance, Epidemiology and End Results Program under contract N01-PC-35136 awarded to the Northern California Cancer Center, contract N01-PC-35139 awarded to the University of Southern California, and contract N02-PC-15105 awarded to the Public Health Institute; and the Centers for Disease Control and Prevention's National Program of Cancer Registries, under agreement #U55/CCR921930-02 awarded to the Public Health Institute. The ideas and opinions expressed herein are those of the author(s) and endorsement by the State of California, Department of Public Health the National Cancer Institute, and the Centers for Disease Control and Prevention or their Contractors and Subcontractors is not intended nor should be inferred. The authors acknowledge the efforts of the Applied Research Program, NCI; the Office of Research, Development and Information, CMS; Information Management Services (IMS), Inc.; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES
  • 1
    American Cancer Society. Cancer Facts and Figures 2008. Atlanta, Ga: American Cancer Society; 2008.
  • 2
    Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2007. CA Cancer J Clin. 2007; 57: 43-66.
  • 3
    Wakabayashi MT, Lin PS, Hakim AA. The role of cytoreductive/debulking surgery in ovarian cancer. J Natl Compr Canc Netw. 2008; 6: 803-810.
  • 4
    Munoz KA, Harlan LC, Trimble EL. Patterns of care for women with ovarian cancer in the United States. J Clin Oncol. 1997; 15: 3408-3415.
  • 5
    Harlan LC, Clegg LX, Trimble EL. Trends in surgery and chemotherapy for women diagnosed with ovarian cancer in the United States. J Clin Oncol. 2003; 21: 3488-3494.
  • 6
    [No authors listed] Screening, treatment, and follow-up. NIH Consensus Development Panel on Ovarian Cancer. JAMA. 1995; 273: 491-497.
  • 7
    McGowan L. Patterns of care in carcinoma of the ovary. Cancer. 1993; 71: 628-633.
  • 8
    McGowan L, Lesher LP, Norris HJ, et al. Misstaging of ovarian cancer. Obstet Gynecol. 1985; 65: 568-572.
  • 9
    Nguyen HN, Averette HE, Hoskins W, et al. National survey of ovarian carcinoma. Part V. The impact of physician's specialty on patients' survival. Cancer. 1993; 72: 3663-3670.
  • 10
    Eisenkop SM, Spirtos NM, Montag TW, et al. The impact of subspecialty training on the management of advanced ovarian cancer. Gynecol Oncol. 1992; 47: 203-209.
  • 11
    Earle CC, Schrag D, Neville BA, et al. Effect of surgeon specialty on processes of care and outcomes for ovarian cancer patients. J Natl Cancer Inst. 2006; 98: 172-180.
  • 12
    Wennberg J, Gittelsohn J. Small area variations in health care delivery. Science. 1973; 182: 1102-1108.
  • 13
    Fisher ES, Wennberg DE, Stukel TA, et al. The implications of regional variations in Medicare spending. Part 1: the content, quality, and accessibility of care. Ann Intern Med. 2003; 138: 273-287.
  • 14
    Fisher ES, Wennberg DE, Stukel TA, et al. The implications of regional variations in Medicare spending. Part 2: health outcomes and satisfaction with care. Ann Intern Med. 2003; 138: 288-298.
  • 15
    RiesLAG, MelbertD, KrapchoM, et al., eds. SEER Cancer Statistics Review, 1975-2005. Bethesda, Md: National Cancer Institute; 2008.
  • 16
    Zippin C, Lum D, Hankey BF. Completeness of hospital cancer case reporting from the SEER Program of the National Cancer Institute. Cancer. 1995; 76: 2343-2350.
  • 17
    Nattinger AB, McAuliffe TL, Schapira MM. Generalizability of the Surveillance, Epidemiology, and End Results Registry population: factors relevant to epidemiologic and health care research. J Clin Epidemiol. 1997; 50: 939-945.
  • 18
    Potosky AL, Riley GF, Lubitz JD, et al. Potential for cancer related health services research using a linked Medicare-tumor registry database. Med Care. 1993; 31: 732-748.
  • 19
    WennbergJE, CooperMM, eds. The Dartmouth Atlas of Health Care. Chicago, Ill: American Hospital Association; 1996.
  • 20
    Klabunde CN, Warren JL, Legler JM. Assessing comorbidity using claims data: an overview. Med Care. 2002; 40: IV-26-IV-35.
  • 21
    Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992; 45: 613-619.
  • 22
    Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987; 40: 373-383.
  • 23
    US Department of Health and Human Services. Area Resource File (ARF). In: Health Resources and Services Administration, Baby Oral Health Program, eds. Rockville, Md: Health Resources and Services Administration, US Department of Health and Human Services; 2006.
  • 24
    Polsky D, Armstrong KA, Randall TC, et al. Variation in chemotherapy utilization in ovarian cancer: the relative contribution of geography. Health Serv Res. 2007; 41: 2201-2218.