Annals of Surgical Oncology Cite Track
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gregorio, D. I.
Right arrow Articles by Zarfos, K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gregorio, D. I.
Right arrow Articles by Zarfos, K.
Related Collections
Right arrow Chemotherapy
Right arrow Radiation Therapy
Right arrow Surgery
Annals of Surgical Oncology 8:844-849 (2001)
© 2001 Society of Surgical Oncology

ORIGINAL ARTICLES

Geographical Differences in Primary Therapy for Early-Stage Breast Cancer

David I. Gregorio, PhD, MS, Martin Kulldorff, PhD, Leah Barry, MPH, Holly Samocuik, BS and Kristin Zarfos, MD

From the Division of Epidemiology & Biostatistics, Department of Community Medicine (DIG, MK, LB, HS) and Department of Surgery (KZ), University of Connecticut School of Medicine, Farmington, Connecticut.

Correspondence: Address correspondence and reprint requests to: David I. Gregorio, Department of Community Medicine, University of Connecticut School of Medicine, Farmington, CT 06030-6205; Fax: 860-679-5464; E-mail: gregorio{at}nso.uchc.edu


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Background: Breast-conserving surgery may not be uniformly available to all women. We evaluated geographical differences across Connecticut in the proportions of cases with early stage breast cancer treated by partial mastectomy (PM). We also looked at geographical variation in PM with axillary lymph node dissection (AND) and PM with adjuvant radiotherapy (RAD).

Methods: The Connecticut Tumor Registry identified 9106 cases of early disease for 1991 to 1995. Latitude-longitude coordinates for place of residence at diagnosis and initial form of therapy were available for 8795 records. A spatial scan statistic was used to detect geographical differences in treatment rates across the state.

Results: A total of 57.7% of early breast cancer cases were treated by PM. Women living around New Haven were less likely than others to be treated in that manner (relative risk [RR] = .86; P = .0001), whereas those living around Norwalk were more likely (RR = 1.26; P = .0001). PM with AND, relative to PM alone, was reported less often for cases from a large area of southwestern Connecticut (RR = .89; P = .0001), but more often for those in north central Connecticut (RR = 1.13; P = .0001). PM with RAD, relative to PM alone, was less common for cases around Danbury (RR = .40; P = .0001) but more common among cases around Hartford (RR = 1.14; P = .0001).

Conclusions: Geographical analysis is a way for physicians and health officials to identify groups of women who may not yet benefit from preferred surgical procedures.

Key Words: Early-stage breast cancer • Geographical differences • Connecticut • Breast-conserving surgery


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Early-stage breast cancer1 can be treated conservatively, with excellent results.2–7 Partial mastectomy (PM) combined with axillary lymph node dissection (AND) or adjuvant radiotherapy (RAD) are considered preferred treatments.2–4,8 By 1995, nearly three fifths of all US cases classified as early disease were treated by PM. Approximately half of those procedures included AND, whereas one in five breast-conserving surgeries was combined with RAD.9 It has been noted that treatment by PM is more common in the presence of stage I disease and among patients 50 years of age or younger, those from high socioeconomic standing, those treated at university hospitals, and those advised by physicians about the availability of the procedure.9–11

Noteworthy variation in the geography of breast-conserving surgery, not readily attributable to patient or disease characteristics, has been reported.12 Rushton and West13 reported sizable variation in mastectomy rates for localized breast cancer across Iowa, as did Guadagnoli et al.14 and Nattinger et al.15 across regions of the United States. Polednak16 found rates of breast-conserving surgery to be high among Connecticut residents who resided near a particular hospital.

The effort toward consistent delivery of effective and desired breast cancer therapies across communities has emerged as an important tertiary cancer control objective.17 Identifying areas of geographical difference in patterns of primary surgical therapy can advance this objective by alerting cancer control specialists to opportunities for patient and provider education. This analysis uses a fine geographical scale to evaluate treatment differences among women across Connecticut with early-stage breast cancer. Geographical differences in use of PM, in relation to subcutaneous, total (simple), modified radical, and radical mastectomies were examined. In addition, we evaluated geographical differences among cases in use of PM with AND or PM with RAD (relative to PM alone).


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
From 1991 to 1995, the Connecticut Tumor Registry identified 9106 incident female breast tumors, according to the International Classification of Diseases (9th revision, clinical modification, code 174; stage I or II based on American Joint Committee on Cancer definitions1). We determined the approximate latitude-longitude coordinates (±100 m) for 8795 records (96.6%), by comparing Connecticut Tumor Registry information on addresses where women lived when their diseases were diagnosed to geographically referenced street files.18,19 Assignment of coordinates was not possible for 311 records because a post office box was listed or the reported address was incomplete or unrecognizable by address-matching software.

Among 8795 early-stage breast cancer records available for analysis, 5075 women (57.7%) were treated by PM. The remainder of records noted more extensive surgeries (i.e., subcutaneous, total (simple), modified radical, or radical mastectomies). Of those treated by PM, AND was reported for 3417 individuals (67.3%). Such information regarding adjuvant therapy is generally regarded as accurate and complete.20 PM with RAD was noted on 3056 records (60.2%). Although accurate, tumor registry information regarding RAD may have undercounted its use in outpatient settings.20

The spatial scan statistic,21 calculated with SaTScan software,22 was used to identify places within Connecticut where treatment differed (either higher or lower) from the statewide experience. By applying a large number of circles of varying size and location, the scan statistic identified the proportion of cases within and outside a particular location treated in a particular manner. Monte Carlo hypothesis testing designated circles where proportions were significantly greater or less than expected.23 The spatial scan statistic makes no a priori assumption about the number of circles to be identified, nor about their size or location. Separate analyses evaluated geographical variation in the proportion of cases treated by (1) PM versus other surgery, (2) PM with AND versus PM alone, or (3) PM with RAD versus PM alone. For every circle identified by the spatial scan statistic, we mapped its approximate location, reported the proportion of affected cases within it, the probability of treatment within relative to outside the identified location (i.e., relative risk estimate), and significance of that estimate, adjusted for the multiple testing inherent in our search.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
The spatial scan statistic identified 12 locations where the proportion of cases treated by PM differed from the statewide experience; at 2 locations, the observed proportions differed significantly from chance.

The most likely location where use of PM differed from other areas of the state was in south central Connecticut (designated as area A in Fig. 1 and Table 1), which includes the cities of New Haven and Bridgeport. Among women with early breast cancer who resided there, only 49.6% were treated by PM. By comparison, 61.0% of cases diagnosed outside of area A were treated in that manner. The probability that women within area A with early breast cancer would receive a PM was estimated to be .86, relative to women residing elsewhere in the state when diagnosed (P = .0001).



View larger version (30K):
[in this window]
[in a new window]
  		FIG. 1. Geographical differences in rates of partial mastectomy (PM) for early stage breast cancer, Connecticut, 1991 to 1995 (8795 cases). Statewide rate of PM = 57.7%; significant P values for greater ({image}) or less ({circ}) than the statewide rate.

 

View this table:
[in this window]
[in a new window]
  		TABLE 1. Selected attributes (%) of early-stage breast cancer cases from Connecticut, 1991–1995
 
A second location around the city of Norwalk (area B) was identified as a place where the use of PM was significantly greater than elsewhere. Within that location, 72.7% of cases were treated in that manner, compared with 56.4% of cases that occurred around Connecticut. The probability of PM treatment for women with early disease who resided in area B was estimated to be 1.26 (P = .0001). Ten additional locations (areas C though L) were identified as having different, but not significant, patterns of PM treatment (Table 2).


View this table:
[in this window]
[in a new window]
  		TABLE 2. Locations with significantly different treatment rates for early-stage breast cancer cases, Connecticut, 1991–1995
 
Geographical differences among early-stage breast cancer cases in use of PM with AND were noted. Statewide, 67.3% of PM cases underwent the procedure. The spatial scan statistic identified a location in Southwestern Connecticut (area M in Fig. 2) that extends from the New York State border to the cities of New Haven and Waterbury in the east, where only 59.6% of cases received PM with AND. Outside that location, PM with AND was reported for 74.7% of cases. We estimated the probability of PM with AND among cases originating within area M to be .89 (P = .0001), relative to cases diagnosed elsewhere. However, cases from north central Connecticut (area N), including the cities of Hartford, Manchester, Willimantic, and Norwich, were significantly more likely than others to be treated in this manner. Whereas 75.8% of cases from area N underwent PM with AND, only 61.8% of cases outside that location underwent the procedure. We estimated the probability that early breast cancer cases from area N would be treated by PM with AND to be 1.13 (P = .0001) relative to elsewhere around the state. A third location, Area O, was identified, but the observed difference at that location did not achieve statistical significance.



View larger version (51K):
[in this window]
[in a new window]
  		FIG. 2. Geographical differences in rates of partial mastectomy with axillary lymph node dissections (PM with AND) among early-stage breast cancer cases, Connecticut, 1991 to 1995 (5075 cases). Statewide rate of PM with AND = 67.3%; significant P value for rates greater ({image}) or less ({circ}) than the statewide rate.

 
The spatial scan statistic identified three locations where use of PM with RAD differed from the statewide rate of 60.2%. Cases originating within area P around Danbury were significantly less likely to receive PM with RAD. Only one in four cases was treated in this manner. The probability a case would receive PM with RAD, relative to other locations around the state, was .40 (P = .0001). Area Q in central Connecticut, by comparison, was a location where PM with RAD was significantly more likely than elsewhere (70.2% vs. 54.4%, respectively). There, the probability of PM with RAD was 1.14 (P = .0001). Within area R, that proportion was increased but did not achieve statistical significance (Fig. 3).



View larger version (71K):
[in this window]
[in a new window]
  		FIG. 3. Geographical differences in rates of partial mastectomy with adjuvant radiotherapy (PM with RAD) among early-stage breast cancer cases, Connecticut, 1991 to 1995 (4935 cases). Statewide rate of PM with RAD = 60.2%; significant P value for rates greater ({image}) or less ({circ}) than the statewide rate.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Advances in breast cancer therapy stem from knowledge derived from clinical trials, experience accrued by practitioners, and preferences expressed by patients. How physicians assist patients in evaluating treatment choices is enormously relevant to evaluating the effectiveness of available therapies and improving disease management. Yet there is painfully little known about the factors that lead physicians and their patients to make clinical decisions under conditions of uncertainty.24 A full understanding of breast cancer treatment decisions will depend on additional information. Qualitative data on physician-patient interaction may illuminate the various factors that influence decision-makers’ judgments.25,26 Physicians’ beliefs and their professional commitments to organizational protocols may influence their presentation of options to patients.27,28 Fear, peer opinion, and prior experiences could be important factors affecting the patients’ preferences for therapy.29 Last, the availability of adjuvant services (e.g., RAD oncology, reconstructive surgery, and so on) is thought to influence decisions about primary therapy.13,30

Consistent with other studies, we found geographical variation in use of breast-conserving therapy for women with early-stage disease.12–16 We extended that information by comparing treatment of women according to exact geographical coordinates for their place of residence at the time of diagnosis. With one exception,16 prior analyses were based on a great level of data aggregation (i.e., region of the country, state, county).

It is not likely that findings reported here were confounded by disease characteristics, inasmuch as there are no reliable data suggesting that such factors differ geographically across the state. Nonetheless, data were stratified according to disease stage (I or II), tumor size (<1 to 20 mm vs. 21 to 50 mm), nodal status (+/-), estrogen receptor status (+/-/unknown), and race/ethnicity (white vs. nonwhite) and analyzed separately. Results of those procedures (not shown) were essentially unchanged; locations with treatment variation reported here were consistently evident in the stratified analyses.

Geographical analysis such as this can be instructive to physicians, patients, and health administrators regarding contextual differences in the provision of care that affects resource allocation, expenditures, outcomes, and patient satisfaction. Consideration of where treatment occurs, by whom, and for what types of cases and patients can reveal why geographical variation occurs and what steps are needed (and likely to be effective) in increasing the consistency of cancer services across communities. The capacity to locate the approximate place of residence of cases allows more precise designation of small areas so that cancer control strategies can effectively address local issues when fashioning community-based programs to encourage breast-conserving therapies.


   Acknowledgments
 
Supported by the American Cancer Society Institutional Grant Award ACSIN152M-151 and the Connecticut Department of Public Health/Centers for Disease Control Award 99-256.

Received for publication March 26, 2001. Accepted for publication August 7, 2001.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Lazovich D, White E, Thomas DB, et al. Underutilizing of breast-conserving surgery and radiation therapy among women with stage I or II breast cancer. JAMA 1991; 266: 3433–8.[Abstract]
  2. NIH consensus conference. Treatment of early-stage breast cancer. JAMA 1991; 265: 391–5.[CrossRef][Medline]
  3. Early Breast Cancer Trialists’ Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic or immune therapy: 133 randomized trials involving 31,000 recurrences and 24,000 deaths among 75,000 women. Lancet 1992; 339: 1–15, 71–85.[Medline]
  4. Bland KI, Scott-Conner CE, Menck H, Winchester DP. Axillary dissection in breast-conserving surgery for stage I and II breast cancer: a National Cancer Data Base study of patterns of omission and implications for survival. J Am Coll Surg 1999; 188: 586–95.[CrossRef][Medline]
  5. Henderson IC. Breast cancer.In: Murphy GP, Lawrence W, Lenhard RE, eds. American Cancer Society Textbook of Clinical Oncology. 2nd ed. Atlanta: American Cancer Society, 1995: 198–219.
  6. Kokubo M, Mitsumori M, Ishikura S, et al. Results of breast-conserving therapy for early stage breast cancer: Kyoto University experiences. Am J Clin Oncol 2000; 23: 499–505.[Medline]
  7. van Dongen JA, Voogd AC, Fentiman IS, et al. Long-term results of a randomized trial comparing breast-conserving therapy with mastectomy: EORTC 10801 trial. J Natl Cancer Inst 2000; 92: 1143–50.[Abstract/Free Full Text]
  8. Cooper MR. Mastectomy from the perspective of a medical oncologist. Cancer 1994; 74; 362–5.[CrossRef][Medline]
  9. Bland KI, McGinnis LS, Morrow M, Eyer H. Clinical highlights from the National Cancer Data Base, 2000. CA Cancer J Clin 2000; 50: 171–83.[Abstract]
  10. Foster RS, Farwell ME, Costanza MC. Breast conserving surgery for breast cancer: patterns of care in a geographic region and estimation of potential applicability. Ann Surg Oncol 1995; 2: 275–80.[Abstract]
  11. Guadagnoli E, Weeks JC, Shapiro CL, et al. Use of breast-conserving surgery for treatment of stage I and stage II breast cancer. J Clin Oncol 1998; 16: 101–6.[Abstract/Free Full Text]
  12. Samet JM, Hunt WC, Farrow DC. Determinants of receiving breast-conserving surgery. The Surveillance, Epidemiology, and End Results Program, 1983–1986. Cancer 1994: 73: 2344–51.[CrossRef][Medline]
  13. Rushton G, West M. Women with localized breast cancer selecting mastectomy treatment, Iowa 1991–1996. Publ Health Rep 1999; 114: 370–1.[Medline]
  14. Guadagnoli E, Weeks JC, Shapiro CL, et al. Use of breast-conserving surgery for treatment of stage I and stage II breast cancer. J Clin Oncol 1998; 16: 101–6.
  15. Nattinger AB, Gottlieb MS, Veum J, et al. Geographic variation in the use of breast-conserving treatment for breast cancer. N Engl J Med 1992; 326: 1102–7.[Abstract]
  16. Polednak AP. Predictors of breast-conserving surgery in Connecticut, 1990–1992. Ann Surg Oncol 1997; 4: 259–63.[Abstract]
  17. Breast Cancer Progress Review Group. Charting the course: priorities for breast cancer research. Chapter 6: treatment. Available at: http://osp.nci.nih.gov/PRGReports/BPRGReport/bprgtreatment.htm Accessed November 17, 1999.
  18. Caliper Corporation. Maptitude User’s Guide, Version 4.1. Newton, MA: Caliper Corporation, 1997.
  19. Expedia Streets98. Redmond, WA: Microsoft Corporation, 1988.
  20. Shambaugh E. Stat bite: accuracy of SEER data. J Natl Cancer Inst 1991; 83: 1369.[Free Full Text]
  21. Kulldorff M. A spatial scan statistic. Commun Stat Theory Methods 1997; 26: 1481–96.[CrossRef]
  22. Kulldorff M, Rand K, Gherman G, et al. SaTScan, Version 2.1 Software for the Spatial and Space-Time Scan Statistics. Bethesda, MD: National Cancer Institute, 1998.
  23. Dwass M. Modified randomization tests for nonparametric hypotheses. Ann Math Stat 1957; 28: 181–7.
  24. Eisenberg JM. Doctor’s Decisions and the Cost of Medical Care. Ann Arbor, MI: Health Administration Press, 1986.
  25. Gregorio DI. Distinguishing styles of practice among physicians who treat breast cancer. Cancer Strategy 1999; 1: 11–5.
  26. Farrow DC, Hunt WC, Samet JM. Geographic variation in the treatment of localized breast cancer. N Engl J Med 1992; 326: 1097–101.[Abstract]
  27. Kotwall CA, Covington DL, Rutledge R, et al. Patient, hospital and surgeon factors associated with breast conservation surgery: a statewide analysis in North Carolina. Ann Surg 1996; 224: 419–29.[CrossRef][Medline]
  28. Lee-Feldstein A, Anton-Culver H, Feldstein PJ. Treatment differences and other prognostic factors related to breast cancer survival: delivery systems and medical outcomes. JAMA 1994; 271; 1163–8.[Abstract]
  29. Ward S, Heidrich S, Wolberg W. Factors women take into account when deciding upon type of surgery for breast cancer. Cancer Nurs 1989; 12: 344–51.[Medline]
  30. Lutz ST, Rinker LH. The surgeon—not geography—influences whether women have lumpectomies vs. mastectomies. Presented to the annual meeting of the Radiological Society of North America, November, 2000.



This article has been cited by other articles:


Home page
 JNCI J Natl Cancer InstHome page
A. Govindarajan, N. G. Coburn, A. Kiss, L. Rabeneck, A. J. Smith, and C. H. L. Law
Population-based assessment of the surgical management of locally advanced colorectal cancer.
J Natl Cancer Inst, October 18, 2006; 98(20): 1474 - 1481.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gregorio, D. I.
Right arrow Articles by Zarfos, K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gregorio, D. I.
Right arrow Articles by Zarfos, K.
Related Collections
Right arrow Chemotherapy
Right arrow Radiation Therapy
Right arrow Surgery

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS