Multiple practice standards and treatment guidelines have been developed for the management of patients with breast carcinoma.1–5 This has been driven by the high incidence rate of breast carcinoma, documentation of significant differences in the use of breast carcinoma treatments, and the fact that breast carcinoma is a major health concern for nearly all women. In general, the goal of practice standards is to improve the quality of care by promoting the use of appropriate therapy and more consistent treatments for all patients with breast carcinoma. The completion of breast-conserving therapy (BCT) for patients with early-stage breast carcinoma is among the outcomes identified that gauge quality of care.6–8 Despite multiple, prospective, randomized trials, some of which now have 10–15 years of follow-up, that have demonstrated equivalent survival from BCT and mastectomy,9–13 the use of BCT has been less than anticipated. After dissemination of the proceedings of the 1990 National Cancer Institute (NCI)/National Institutes of Health (NIH) consensus development conference, which concluded that BCT was an appropriate and preferable method of treatment for the majority of women with Stage I and II breast carcinoma,1 there has been only a modest increase in BCT use14 that still varies depending on patient and hospital characteristics.15, 16
In 1992, the American College of Surgeons (ACOS), the American College of Radiology (ACR), the College of American Pathologists, and the Society of Surgical Oncology collaborated in establishing standards for BCT.17 These standards identified critical elements for appropriately selecting and managing patients who receive BCT. The ACR Patterns of Care Studies (PCS) and the ACOS Commission on Cancer (COC), both of which have surveyed breast carcinoma treatment methods,18, 19 launched a joint study to establish practice patterns for local therapy in patients with early-stage breast carcinoma since the dissemination of the 1992 BCT standards. Specifically, this study sought to determine, in a large population of eligible patients with breast carcinoma, the rate of BCT use versus mastectomy and the degree of adherence to the 1992 multispecialty-endorsed BCT standards 2 years after their dissemination.
The patterns of local therapy in the population were reported previously.20 A breast-conservation surgery (BCS) rate of 42.6% was documented with significant variation noted by hospital, patient, and tumor factors.20 Multivariate analysis demonstrated that treatment in the Northeast region, clinical T1 tumor, and absence of an extensive intraductal component were the strongest predictors for undergoing BCS. Radiation therapy (RT) was given to 86% of patients who underwent BCS. Age < 70 years was the strongest predictor for receiving RT. The objectives of this report were to determine 1) the penetration and rate of adherence to the 1992 standards for BCT 2 years after publication and 2) whether compliance to the standards for BCT varies by patient and hospital factors associated with the quantity of BCT performed in the entire population.
MATERIALS AND METHODS
Representatives from the ACR and the ACOS met and, through consensus, developed a study specific questionnaire to achieve the goals of the study. The questionnaire consisted of 72 questions, including general disease information (laterality, histology, etc.), diagnostic and staging procedures, disease-directed surgeries, tumor markers/prognostic indicators, American Joint Committee on Cancer clinical and pathologic stages, pathologic evaluation, radiation oncology worksheet, and follow-up.
An invitation to participate in the study was sent to 2000 hospitals in the United States, including 1400 hospitals with cancer programs accredited by the ACOS COC. Each program was asked to submit data on the first 25 consecutive patients with of Stage I or II breast carcinoma who were treated in 1994. Our assumption was that at least 2 years were necessary after the initial publication of the standards in 199217 for penetration into medical practice. Hospital cancer registrars voluntarily submitted data. For this study, a representative of the institution's radiation oncology department completed the radiation oncology worksheet portion of the questionnaire. The patients who were included had to be diagnosed and had to receive their initial course of treatment at the participating institution. A total of 842 hospitals submitted data. Ninety-three percent of hospitals had ACOS-approved cancer programs.
The submitted data yielded 17,931 patients with breast carcinoma who were diagnosed in 1994. A total of 16,643 patients with pathologic Stage I–II breast carcinoma were analyzed for this report. All disease stages, tumor sizes, and lymph node status reported herein are pathologic and according to the American Joint Committee on Cancer staging manual, fourth edition.21 A total of 1288 patients were excluded from this analysis. In 984 patients (5.4%), the submitted information was incomplete and did not allow for reliable pathologic staging. Another 304 patients (1.7%) were excluded because they would not be considered appropriate candidates for BCT based on the standards:17 Six patients had Stage III disease, 260 patients had were T3 tumors, and 38 patients had undergone surgery other than lumpectomy or mastectomy.
Definition of Variables for Analysis
For the frequency of adherence to standards analyses, the patients in the black and Hispanic racial groups were combined into one group, black and Hispanic. An other race group included Asians, Pacific Islanders, and American Indians/Eskimos.
The insurance carriers were divided into two large groups: private, which included health maintenance organizations (HMOs), preferred provider organizations (PPOs) and private insurance companies; and government, which included federally and state-funded programs. The federally funded programs included Medicare with or without supplement as well as Veterans Administration, CHAMPUS, military, and Indian health services. State-funded insurance included Medicaid, welfare, and a small number of uninsured patients.
Hospitals were grouped into the same six geographic regions for this analysis (Table 1) that have been used by the ACOS for previous studies.19 Hospital cancer programs were grouped into three categories—comprehensive community, community, or teaching—based on ACOS COC definitions. Comprehensive community cancer programs have board-certified oncology specialists, a full range of diagnostic and treatment services on site or by referral, and weekly cancer conferences. The community cancer program variable includes three categories: the Community Hospital Cancer Program (CHCP), the Hospital-Associated Cancer Program (HACP), and the Integrated Cancer Program (ICP). A CHCP has less than a full range of diagnostic and treatment services and usually conducts cancer conferences less often than weekly. An HACP has a limited range of diagnostic and cancer treatment services available on site, and cancer conferences are held monthly. An ICP hospital offers one treatment modality and forms partnerships with another hospital facility to provide other services. The teaching group included Teaching Hospital Cancer Programs (THCPs) and NCI-designated Comprehensive Cancer Programs (NCIPs). The THCPs are based in institutions with at least four full residency programs and have a full range of board-certified specialty services in oncology diagnosis and treatment as well as weekly cancer conferences. An NCIP program, in addition to the services available at THCPs, conducts both basic and clinical research.
Table 1. Geographic Regions
|Northeast||Maine, Vermont, New Hampshire, Massachusetts, Rhode Island, Connecticut, New York, Pennsylvania, New Jersey|
|South Atlantic||Delaware, District of Columbia, Maryland, West Virginia, Virginia, North Carolina, South Carolina, Georgia, Florida|
|Midwest||Wisconsin, Michigan, Illinois, Indiana, Ohio, Minnesota, North Dakota, South Dakota, Iowa, Nebraska, Kansas, Missouri|
|South||Kentucky, Tennessee, Mississippi, Alabama, Oklahoma, Arkansas, Texas, Louisiana|
|Mountain||Montana, Idaho, Wyoming, Nevada, Utah, Colorado, Arizona, New Mexico|
|Pacific||Washington, Oregon, California, Alaska, Hawaii|
STANDARED FOR BCT
Twenty-one standards identified from the 1992 Standards for Breast-Conservation Treatment17 were evaluated in the following categories.
Mammography (two standards)
The standards state that preoperative mammographic evaluation is necessary to determine whether patients are eligible for BCT, and the size of the mammographic abnormality should be included in the report.17
Labeling of the surgical specimen (three standards)
The excised tissue must be submitted for pathologic examination with appropriate clinical history as well as anatomic site specifications, including laterality, quadrant, and orientation of the specimen with sutures or other markers.17
Content of pathology report (10 standards)
The pathologist includes certain basic data in each surgical pathology report, because a final decision on the type of local therapy usually is not made before the pathology examination is complete.17 The basic data set should include microscopic confirmation of disease, measurement of the size of the carcinoma, histologic type, histologic grade, presence or absence of invasion of lymphatic and blood vessels, presence or absence of macroscopic or microscopic tumor in the margins of excision, and steroid receptor analysis.17
RT (six standards)
BCT is defined by the 1992 standards as BCS followed by RT.17 The elements of recommended radiation techniques include treating each field 5 days per week, treatment planning on a dedicated simulator, documentation that a calculated dose planning (dosimetry) had been performed, dose planning with the use of tissue compensators to improve the homogeneity of dose distribution, and the avoidance of breast bolus to minimize skin toxicity. The 15 standards in the areas of preoperative mammography, surgical specimen labeling, and pathology report content are grouped as perioperative standards.
Standards for Systemic Therapy
Adjuvant systemic therapy was not included in the 1992 BCT standards, because they were concentrated solely on local therapy. However, that same year, the St. Gallen Conference treatment recommendations for adjuvant systemic therapy in patients with primary breast carcinoma were reported.22 The St. Gallen Conference treatment recommendations stated that all patients who are treated for with lymph node positive breast carcinoma should receive some form of systemic therapy.22 Because the use of systemic therapy is considered an indicator of the quality of treatment for patients with early stage breast carcinoma,1, 2, 6, 7 it also was evaluated for comprehensiveness.
The rate of adherence to standards is reported in frequency tables. Scores > 80% were considered consistent with the standard. Chi-square statistics were calculated for comparisons of adherence among patient and hospital variables. For each patient and hospital variable, the percent of standards with significant variation also was calculated. A P value < 0.05 was considered statistically significant. For each frequency table, the variable that predicted for more BCS or RT on multivariate analysis compared with the prior analysis20 is indicated.
For the patients undergoing lumpectomy who were diagnosed in 1994, practice appeared to be consistent with 16 of 22 standards (73%) that were evaluated from the published 1992 Standards for Breast-Conservation Treatment.17 When we examined the pattern of compliance among the patient and hospital variables, we found that significant variation was common based on the type of hospital cancer program and geographic location. In this population, factors that have predicted for more BCT use were not associated with more frequent compliance to standards.
The goal of the Standards for Breast-Conservation Treatment was to promote better and more consistent care of women with breast carcinoma.17 The standards emphasized the necessary collaboration and interdependence of multiple specialists for optimal outcomes for patients who receive BCT. For this reason, specific standards were identified for radiology, surgery, pathology, and radiation oncology practices that were deemed pertinent and essential to successful BCT. Other existing practice guidelines for the treatment of patients with breast carcinoma are more comprehensive, including all aspects of treatment after women are diagnosed with breast carcinoma3 and the management of populations at risk.4 These other guidelines address BCT primarily by defining standards for the selection of appropriate patients based on absolute and relative contraindications.3, 4 In addition to defining patient selection, the BCT standards established guidelines for the process involved in the successful completion of BCT.
For treatment standards to be useful or to have an impact on patient care, they must be both valid and followed. An important component identified for maintaining the quality of practice guidelines has been periodic review and revision to keep pace with emerging research and technologies.23, 24 In 1998, the initial 1992 BCT standards were reviewed, revised, and approved by the same consortium of specialty societies.25 All 21 standards that we evaluated from the 1992 version remained in the updated revision.25 In addition, patient selection, mammography, surgical techniques, pathology report content, and RT technique, as specified in the 1992 BCT standards, were validated further by other comprehensive guidelines3, 4 and by individual specialty reports.26–29
In this study, the adherence to 16 of 21 standards was > 80%, and adherence was > 90% for 10 standards. Although, ideally, the adherence rate should be as close to 100% as possible, the rate that is sufficiently high enough to be considered acceptable has not been well established. Other studies examining adherence to treatment guidelines for patients with breast carcinoma reported compliance ≥ 80% as reflective of care consistent with guidelines.7, 30 Guadagnoli et al. evaluated four standards that were identified from the 1991 NIH consensus statement: 1) RT after lumpectomy, 2) axillary lymph node dissection, 3) chemotherapy for premenopausal patients with positive lymph nodes, and 4) hormone therapy for postmenopausal patients with positive receptor status and positive lymph nodes. The compliance rates for three of the four standards of care studied were > 80% and were considered consistent with the results of the national consensus conference.7 The compliance rates for hormone therapy were considered low at 59% and 63% in Minnesota and Massachusetts, respectively.7 In another study, compliance to practice guidelines for the treatment of patients with breast carcinoma in British Columbia for adjuvant RT, chemotherapy, and tamoxifen was 97%, 96%, and 89%, respectively, for 939 patients: Compliance was considered high or closely matched to treatment guidelines in that study.31 Our study, in comparison, demonstrated similar good compliance, with 85.9% of women receiving adjuvant RT after BCS and 84.1% of patients with positive lymph nodes receiving systemic therapy.
Poor compliance was found to six standards in our study: 1) documentation of the size of the mammography abnormality in the report, 2) labeling the lumpectomy specimen with the affected quadrant of the breast, 3) spatially orienting the lumpectomy specimen with sutures or other markers, 4) documenting the presence or absence of lymphatic and/or vascular invasion in the pathology report, 5) documenting the presence of absence of DCIS in the pathology report, and 6) documenting the macroscopic margin status in the pathology report. In this study, only 47% of mammogram reports from patients who underwent lumpectomy contained a measurement. We examined mammography reports from the patients who underwent mastectomy in this series to determine whether this finding was surgery specific. Even fewer patients who underwent mastectomy had mammogram reports (39.6%) that included a tumor size. This may have been because a high percentage of ultrasound examinations were done at the time of the mammogram, and a tumor size was included in that report. More likely, it reflects that the ACR standard for the performance of diagnostic mammography (2000–2001) and the ACR Illustrated Breast Imaging and Reporting Data Systems Committee do not include abnormality size among the items to be described in the mammography report.28, 32 This is unfortunate, because an assessment of tumor size relative to the patient's breast size is a critical element in determining the patient's suitability for BCS.
Although 98% of lumpectomy specimens were labeled for laterality, only 21.1% had the quadrant of the breast identified, and 67% were oriented spatially with sutures or other markers. The Association of Directors of Anatomic and Surgical Pathologists (ADASP) recommendations for reporting breast carcinoma includes quadrant in the labeling only if the specimen is a mastectomy for the treatment of breast carcinoma.26 Quadrant identification of the lumpectomy specimen is the only standard with poor compliance that deserves rethinking it terms of its overall utility as a standard. In contrast, surgical and pathology groups alike have identified proper orientation of the lumpectomy specimen by the surgeon as an important step for determining adequacy of excision.26, 27
The other three standards with lower compliance all were items to be included in the lumpectomy pathology report. The standard for documenting a macroscopic margin evaluation had 72.4% compliance, documenting the present or absence of lymphatic and/or vascular invasion had a compliance rate of 53.5%, and documenting the presence or absence of DCIS had a compliance rate of 42.2%. Each of these standards also is recommended for inclusion in the pathology report by the ADASP. In addition, other studies have demonstrated that the presence and quantity of lymphatic/vascular invasion and/or DCIS in the lumpectomy specimen can be correlated with outcome after BCT.33, 34
We previously demonstrated that, in this population, certain patient and hospital factors are associated with less use of BCT.20 For the most part, variables that demonstrated significant variation in compliance to the standards were not the variables that predicted for less BCS use in the entire population. Nattinger et al., using SEER data, noted that, during 1990–1995, when the rates of BCT had increased, there also was a 9% increase in inappropriate care.35 In the current study, which was designed specifically to assess compliance to an established breast-conservation standard, we did not see less quality among variables that were associated with greater utilization of BCT.
Significant differences in compliance occurred commonly based on the geographic location of the hospital and the type of cancer program rather than the patient factors of age, race, and payer. When there was significant variation in compliance to the perioperative standards (mammography, surgical specimen, or pathology report content) by cancer programs, community hospitals without comprehensive cancer programs frequently were the least compliant (Table 8). Over 7000 patients, or 42% of the overall study population, were treated in this type of community hospital program and had the lowest percentage of BCS use among the types of cancer programs. Other studies have demonstrated less use of adjuvant therapies based on a hospital's case load of patients with breast carcinoma.30 In the future, hospitals with this type of cancer program should be targeted for distribution of practice standards for the treatment of these patients. Although variation in compliance by the hospital's geographic region also was common for the 15 perioperative standards, one region was not the least compliant more frequently compared with the other regions.
Teaching hospital cancer programs frequently were the least compliant when there was significant variation to the adjuvant RT standards. This was confined to adherence to the techniques for radiation therapy and not in its use. Prior surveys by ACR/PCS in 1983 and 1989 regarding breast RT did not demonstrate significant differences in the process of care or treatment planning for academic facilities versus nonacademic facilities.29, 36 Future PCS surveys should reevaluate for any differences in the delivery of RT by hospital type in view of the findings in this study.
Our study had several limitations. First, it is not known whether all of the standards should be expected to have a similar compliance rate, i.e., a rate that should approach 100%. For instance, standards may have a < 100% compliance rate due to patient refusal or the presence of comorbidity. Our study was not designed to detect legitimate reasons for noncompliance. Second, our population may not reflect practice patterns in the country as a whole. There is an internal bias in this study, because all participating hospital registrars voluntarily submitted data to the study and predominantly were approved by the ACOS COC. We recognize that our study population represented a large convenience sample of patients with early-stage breast carcinoma that allowed us to observe trends in treatment. Finally, the large population size makes relatively small differences in compliance become statistically significant using a 0.05 level. Caution should be used particularly in interpreting small differences between groups or small subsets of the population.
Although multiple guidelines for the treatment of patients with breast carcinoma have been disseminated, there are less data evaluating the penetration, adherence rate, and practical utility of guidelines.7, 31, 37 The current study demonstrates that large-scale evaluation of practice standards is feasible and that 2 years postdissemination seems to be a good evaluation time point, as judged by the high adherence rate to most standards. Evaluation is necessary to demonstrate whether the guideline is having the expected impact and to allow review and revisions as necessary when the prevailing practices do not follow the guideline.2, 3, 24, 38 The goal of these BCT guidelines was to promote quality care and to identify areas in which physician and patient education were necessary for achieving optimal outcomes from BCT.17 We found that, for patients who received BCT, treatment practice was consistent for 73% of the 22 standards that we evaluated 2 years after their dissemination, with poor compliance for 6 standards. The standards with poor compliance represent a call for targeted physician education and reevaluation for improvement. Professional groups may be suited best for outreach to physicians, particularly in smaller community hospital cancer programs and certain geographic locations, to ensure penetration of treatment recommendations. In this study, compliance differences commonly occurred depending on the hospital's geographic location and type of cancer program. Based on our findings, in the future, organizations promulgating standards for the treatment of patients carcinoma should provide focused distribution and evaluation of standards on certain cancer practices and locations to ensure homogenous adherence to quality care.